Re: Dynamic phases proposal

classic Classic list List threaded Threaded
14 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
Robert, Michael, Tibor, let’s continue here (though I asked Infra and it’s
fine that anyone in the community can join our Slack)

On Fri 25 Oct 2019 at 20:01, Stephen Connolly <
[hidden email]> wrote:

> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>
> Thoughts?
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
Robert,

I would be fine splitting out into, pardon the pun, phases:

Phase 1: before and after
Phase 2: priorities
Phase 3: transitional lifecycle

Might have a phase 1.5 of before:* and after:* to catch the start of a
lifecycle and the end of a lifecycle...

On Fri 25 Oct 2019 at 20:30, Stephen Connolly <
[hidden email]> wrote:

> Robert, Michael, Tibor, let’s continue here (though I asked Infra and it’s
> fine that anyone in the community can join our Slack)
>
> On Fri 25 Oct 2019 at 20:01, Stephen Connolly <
> [hidden email]> wrote:
>
>> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>>
>> Thoughts?
>> --
>> Sent from my phone
>>
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

Robert Scholte-8
TLDR: We can do better than, but who is in control? lifecycle-owner, plugin-owner or pom-owner?

I think we all recognize the issues we're trying to solve, but to me this proposal is not the right solution.

In general there are 2 issues:
1. provide a mechanism that makes sure some executions are called even its matching main phase fails.
2. provide a mechanism then ensures the order of executions. 

The problem of issue 1 is described in MNG-5668, but not the final solution. 
MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas stage 2 proposes to give the power to the *pom-owner*. 
Both agree on the same thing: by default these post-phases should be triggered even after failure of the matching main phase. This is actually already expected behavior, so I don't expect real issues when implementing this adjusted behavior.
To me after:integration-test is just an alias for post-integration-test, both should work the same way.

Issue 2 is a more common problem: controlling the order of executions. 
In some cases it is pretty hard or even impossible to get the preferred order. The latter happens when 2 goals of the same plugin must be executed and a goal of another plugin are competing within the same phase.

So let's first take a look at a phase: is there a clear definition?
"A phase is a step in what Maven calls a 'build lifecycle'. The build lifecycle is an ordered sequence of phases involved in building a project".
"Lifecycle phases are intentionally vague, defined solely as validation, testing, or deployment, and they may mean different things to different projects."
Phases are intended to be called from the commandline, and within the pom you define you can control what should happen before or during that phase.

To me changing the content of the <phase>-element is a codesmell as it becomes more than just a phase, and we start programming. Why do we need it?
In the end it is all about ensuring the order of plugin executions. Stage3+4 proposes to give the power to the *pom-owner*, whereas MPLUGIN-350[2] proposes to give this power to the *plugin-owner*.
IIUR Gradle does not have this issue, because their plugins are aware of input and output. They ensure that if the output plugin X is the input of plugin Y, than X is executed before Y. 
And we should do the same. And this comes with benefits: we can decide if executions within a project can be executed in parallel. And the pom stays as clean as it is right now.

In cases when there's a better ownership than the pom-owner, I would prefer to choose that solution. I already notice how people (don't) build up their knowledge regarding poms. The lifecycle-owner and plugin-owner know much better what they're doing.

thanks,
Robert

Some food for thoughts: consider a developer that wants to run up until pre-integration-test, because he wants to bring his system in a certain state so he can work with IDE to do some work.Can we say that If And Only If somebody called the pre-PHASE, there's no reason to end with the post-PHASE?

[1] https://issues.apache.org/jira/browse/MNG-5668
[2] https://issues.apache.org/jira/browse/MPLUGIN-350
On 26-10-2019 14:20:50, Stephen Connolly <[hidden email]> wrote:
On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:

> To avoid confusion, let's call it stages.
>
> Stage 1: Always call post-bound executions (MNG-5665[1])
> Stage 2: before and after
> Stage 3: priorities (MNG-3522[2])
> Stage 4: transitional lifecycle


I have a prototype of stages 1-3 nearly (80%) done... just have to polish
up and validate the bound executions with some tests


>
> For both all you need to start evaluating the value of phase.
> For now we can assume that after:clean is just another label for
> post-clean and will have exactly the same effect.
> MNG-5665 contains a proposal to change the xml, but we shouldn't do that
> (yet). Let's start with a hardcoded list of postphases (or in case a goal
> fails, see if a post-x phase exists). Stage 1 is to make it work, stage 2
> to make it configurable.
> IIRC you cannot ask from inside a Mojo if is was called explicitly or
> because it was bound to a phase, nor can you ask for the value of this
> phase. I kind of like this, plugins shouldn't care about this.
> However, inside Maven it will become important at which phase it is to
> know if there are more executions to call OR create blocks of executions.
> Now it is just a list of executions: loop and fail fast.
>
> thanks,
> Robert
>
> [1] https://issues.apache.org/jira/browse/MNG-5665
> [2] https://issues.apache.org/jira/browse/MNG-3522
> On 25-10-2019 21:33:14, Stephen Connolly
> wrote:
> Robert,
>
> I would be fine splitting out into, pardon the pun, phases:
>
> Phase 1: before and after
> Phase 2: priorities
> Phase 3: transitional lifecycle
>
> Might have a phase 1.5 of before:* and after:* to catch the start of a
> lifecycle and the end of a lifecycle...
>
> On Fri 25 Oct 2019 at 20:30, Stephen Connolly <>
> [hidden email] [mailto:[hidden email]]>
> wrote:
>
> Robert, Michael, Tibor, let’s continue here (though I asked Infra and it’s
> fine that anyone in the community can join our Slack)
>
> On Fri 25 Oct 2019 at 20:01, Stephen Connolly <>
> [hidden email] [mailto:[hidden email]]>
> wrote:
>
> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
>
> Thoughts?
> --
>
> Sent from my phone
> --
>
> Sent from my phone
> --
>
> Sent from my phone

--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
https://github.com/apache/maven/tree/mng-5668-poc is my POC implementation
for anyone interested in trying it out.

Here's a pom that builds with the PoC

<project>
    <modelVersion>4.0.0</modelVersion>
    <groupId>localdomain</groupId>
    <artifactId>foo</artifactId>
    <version>1.0-SNAPSHOT</version>
    <build>
        <plugins>
            <plugin>
                <artifactId>maven-antrun-plugin</artifactId>
                <executions>
                    <execution>
                        <id>1</id>
                        <phase>before:integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <echo message="hi from execution 1"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>2</id>
                        <phase>before:integration-test[1000]</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <echo message="execution 2 happens before 1
even though later in the pom"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>3</id>
                        <phase>after:integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <echo message="execution 3 always executes
if we start the integration-test phase"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>4</id>
                        <phase>integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <fail/><!-- time to fail the build-->
                            </tasks>
                        </configuration>
                    </execution>
                </executions>
            </plugin>
        </plugins>
    </build>
</project>


On Sun, 27 Oct 2019 at 10:55, Robert Scholte <[hidden email]> wrote:

> TLDR: We can do better than, but who is in control? lifecycle-owner,
> plugin-owner or pom-owner?
>
> I think we all recognize the issues we're trying to solve, but to me this
> proposal is not the right solution.
>
> In general there are 2 issues:
> 1. provide a mechanism that makes sure some executions are called even its
> matching main phase fails.
> 2. provide a mechanism then ensures the order of executions.
>
> The problem of issue 1 is described in MNG-5668, but not the final
> solution.
> MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> stage 2 proposes to give the power to the *pom-owner*.
> Both agree on the same thing: by default these post-phases should be
> triggered even after failure of the matching main phase. This is actually
> already expected behavior, so I don't expect real issues when implementing
> this adjusted behavior.
> To me after:integration-test is just an alias for post-integration-test,
> both should work the same way.
>
> Issue 2 is a more common problem: controlling the order of executions.
> In some cases it is pretty hard or even impossible to get the preferred
> order. The latter happens when 2 goals of the same plugin must be executed
> and a goal of another plugin are competing within the same phase.
>
> So let's first take a look at a phase: is there a clear definition?
> "A phase is a step in what Maven calls a 'build lifecycle'. The build
> lifecycle is an ordered sequence of phases involved in building a project".
> "Lifecycle phases are intentionally vague, defined solely as
> validation, testing, or deployment, and they may mean different things to
> different projects."
> Phases are intended to be called from the commandline, and within the pom
> you define you can control what should happen before or during that phase.
>
> To me changing the content of the <phase>-element is a codesmell as it
> becomes more than just a phase, and we start programming. Why do we need it?
> In the end it is all about ensuring the order of plugin executions.
> Stage3+4 proposes to give the power to the *pom-owner*,
> whereas MPLUGIN-350[2] proposes to give this power to the *plugin-owner*.
> IIUR Gradle does not have this issue, because their plugins are aware of
> input and output. They ensure that if the output plugin X is the input of
> plugin Y, than X is executed before Y.
> And we should do the same. And this comes with benefits: we can decide if
> executions within a project can be executed in parallel. And the pom stays
> as clean as it is right now.
>
> In cases when there's a better ownership than the pom-owner, I would
> prefer to choose that solution. I already notice how people (don't) build
> up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> know much better what they're doing.
>
> thanks,
> Robert
>
> Some food for thoughts: consider a developer that wants to run up until
> pre-integration-test, because he wants to bring his system in a certain
> state so he can work with IDE to do some work.Can we say that If And Only
> If somebody called the pre-PHASE, there's no reason to end with the
> post-PHASE?
>
> [1] https://issues.apache.org/jira/browse/MNG-5668
> [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> On 26-10-2019 14:20:50, Stephen Connolly <[hidden email]>
> wrote:
> On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
>
> > To avoid confusion, let's call it stages.
> >
> > Stage 1: Always call post-bound executions (MNG-5665[1])
> > Stage 2: before and after
> > Stage 3: priorities (MNG-3522[2])
> > Stage 4: transitional lifecycle
>
>
> I have a prototype of stages 1-3 nearly (80%) done... just have to polish
> up and validate the bound executions with some tests
>
>
> >
> > For both all you need to start evaluating the value of phase.
> > For now we can assume that after:clean is just another label for
> > post-clean and will have exactly the same effect.
> > MNG-5665 contains a proposal to change the xml, but we shouldn't do that
> > (yet). Let's start with a hardcoded list of postphases (or in case a goal
> > fails, see if a post-x phase exists). Stage 1 is to make it work, stage 2
> > to make it configurable.
> > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > because it was bound to a phase, nor can you ask for the value of this
> > phase. I kind of like this, plugins shouldn't care about this.
> > However, inside Maven it will become important at which phase it is to
> > know if there are more executions to call OR create blocks of executions.
> > Now it is just a list of executions: loop and fail fast.
> >
> > thanks,
> > Robert
> >
> > [1] https://issues.apache.org/jira/browse/MNG-5665
> > [2] https://issues.apache.org/jira/browse/MNG-3522
> > On 25-10-2019 21:33:14, Stephen Connolly
> > wrote:
> > Robert,
> >
> > I would be fine splitting out into, pardon the pun, phases:
> >
> > Phase 1: before and after
> > Phase 2: priorities
> > Phase 3: transitional lifecycle
> >
> > Might have a phase 1.5 of before:* and after:* to catch the start of a
> > lifecycle and the end of a lifecycle...
> >
> > On Fri 25 Oct 2019 at 20:30, Stephen Connolly <>
> > [hidden email] [mailto:[hidden email]
> ]>
> > wrote:
> >
> > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> it’s
> > fine that anyone in the community can join our Slack)
> >
> > On Fri 25 Oct 2019 at 20:01, Stephen Connolly <>
> > [hidden email] [mailto:[hidden email]
> ]>
> > wrote:
> >
> > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> >
> > Thoughts?
> > --
> >
> > Sent from my phone
> > --
> >
> > Sent from my phone
> > --
> >
> > Sent from my phone
>
> --
> Sent from my phone
>
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
On Tue 12 Nov 2019 at 07:34, Robert Scholte <[hidden email]> wrote:

> This is not just MNG-5668, but also contains several non-existing issues,
> that should be mentioned explicitly as they will have huge impact:
>
> - support before:/after: prefix for phase-binding
>
> - introduce priority
> - reduce phases (this one hasn't been implemented, but seems to be the
> reason behind before:/after:)


All detailed in the proposal on the wiki:
https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases

Reducing phases would be a big change and not before 4.x at least (maybe
5.x more realistically... at least we’d need to deprecate the phases for a
good while before removing any)


>
> I would like see separate branches for all of them, as they all have their
> own discussion.


The whole point of a PoC is the get feedback. I don’t see utility in
separate branches as they are all touching the same code.

Once we get feedback we can decide where we want to go from there.


>
> Robert
> On 11-11-2019 20:31:44, Stephen Connolly <[hidden email]>
> wrote:
> https://github.com/apache/maven/tree/mng-5668-poc is my POC implementation
> for anyone interested in trying it out.
>
> Here's a pom that builds with the PoC
>
>
> 4.0.0
> localdomain
> foo
> 1.0-SNAPSHOT
>
>
>
> maven-antrun-plugin
>
>
> 1
> before:integration-test
>
> run
>
>
>
>
>
>
>
>
> 2
> before:integration-test[1000]
>
> run
>
>
>
>
>
>
>
>
> 3
> after:integration-test
>
> run
>
>
>
>
>
>
>
>
> 4
> integration-test
>
> run
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
>
> > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > plugin-owner or pom-owner?
> >
> > I think we all recognize the issues we're trying to solve, but to me this
> > proposal is not the right solution.
> >
> > In general there are 2 issues:
> > 1. provide a mechanism that makes sure some executions are called even
> its
> > matching main phase fails.
> > 2. provide a mechanism then ensures the order of executions.
> >
> > The problem of issue 1 is described in MNG-5668, but not the final
> > solution.
> > MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> > stage 2 proposes to give the power to the *pom-owner*.
> > Both agree on the same thing: by default these post-phases should be
> > triggered even after failure of the matching main phase. This is actually
> > already expected behavior, so I don't expect real issues when
> implementing
> > this adjusted behavior.
> > To me after:integration-test is just an alias for post-integration-test,
> > both should work the same way.
> >
> > Issue 2 is a more common problem: controlling the order of executions.
> > In some cases it is pretty hard or even impossible to get the preferred
> > order. The latter happens when 2 goals of the same plugin must be
> executed
> > and a goal of another plugin are competing within the same phase.
> >
> > So let's first take a look at a phase: is there a clear definition?
> > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > lifecycle is an ordered sequence of phases involved in building a
> project".
> > "Lifecycle phases are intentionally vague, defined solely as
> > validation, testing, or deployment, and they may mean different things to
> > different projects."
> > Phases are intended to be called from the commandline, and within the pom
> > you define you can control what should happen before or during that
> phase.
> >
> > To me changing the content of the -element is a codesmell as it
> > becomes more than just a phase, and we start programming. Why do we need
> it?
> > In the end it is all about ensuring the order of plugin executions.
> > Stage3+4 proposes to give the power to the *pom-owner*,
> > whereas MPLUGIN-350[2] proposes to give this power to the *plugin-owner*.
> > IIUR Gradle does not have this issue, because their plugins are aware of
> > input and output. They ensure that if the output plugin X is the input of
> > plugin Y, than X is executed before Y.
> > And we should do the same. And this comes with benefits: we can decide if
> > executions within a project can be executed in parallel. And the pom
> stays
> > as clean as it is right now.
> >
> > In cases when there's a better ownership than the pom-owner, I would
> > prefer to choose that solution. I already notice how people (don't) build
> > up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> > know much better what they're doing.
> >
> > thanks,
> > Robert
> >
> > Some food for thoughts: consider a developer that wants to run up until
> > pre-integration-test, because he wants to bring his system in a certain
> > state so he can work with IDE to do some work.Can we say that If And Only
> > If somebody called the pre-PHASE, there's no reason to end with the
> > post-PHASE?
> >
> > [1] https://issues.apache.org/jira/browse/MNG-5668
> > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > On 26-10-2019 14:20:50, Stephen Connolly
> > wrote:
> > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> >
> > > To avoid confusion, let's call it stages.
> > >
> > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > Stage 2: before and after
> > > Stage 3: priorities (MNG-3522[2])
> > > Stage 4: transitional lifecycle
> >
> >
> > I have a prototype of stages 1-3 nearly (80%) done... just have to polish
> > up and validate the bound executions with some tests
> >
> >
> > >
> > > For both all you need to start evaluating the value of phase.
> > > For now we can assume that after:clean is just another label for
> > > post-clean and will have exactly the same effect.
> > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> that
> > > (yet). Let's start with a hardcoded list of postphases (or in case a
> goal
> > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> stage 2
> > > to make it configurable.
> > > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > > because it was bound to a phase, nor can you ask for the value of this
> > > phase. I kind of like this, plugins shouldn't care about this.
> > > However, inside Maven it will become important at which phase it is to
> > > know if there are more executions to call OR create blocks of
> executions.
> > > Now it is just a list of executions: loop and fail fast.
> > >
> > > thanks,
> > > Robert
> > >
> > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > On 25-10-2019 21:33:14, Stephen Connolly
> > > wrote:
> > > Robert,
> > >
> > > I would be fine splitting out into, pardon the pun, phases:
> > >
> > > Phase 1: before and after
> > > Phase 2: priorities
> > > Phase 3: transitional lifecycle
> > >
> > > Might have a phase 1.5 of before:* and after:* to catch the start of a
> > > lifecycle and the end of a lifecycle...
> > >
> > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > [hidden email] [mailto:
> [hidden email]
> > ]>
> > > wrote:
> > >
> > > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> > it’s
> > > fine that anyone in the community can join our Slack)
> > >
> > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > [hidden email] [mailto:
> [hidden email]
> > ]>
> > > wrote:
> > >
> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > >
> > > Thoughts?
> > > --
> > >
> > > Sent from my phone
> > > --
> > >
> > > Sent from my phone
> > > --
> > >
> > > Sent from my phone
> >
> > --
> > Sent from my phone
> >
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

Robert Scholte-8
The name of the branch contains MNG-5668, but it contains much more.
I'd likely lead to comments like "great", without being explicit saying which part(s).
I am aware there's all proposals touch the same code, but can be released isolated from each other.
e.g. if the enums-value are changed to "pre-" and "post-" it should work for the existing phases, which means we could already use it quite soon (still need to test it myself, though)
I also want to provide a counter proposal, but that takes time and for me there are other issues more important.

My biggest fear is that this will result in an All-Or-Nothing, and I like to prevent that. If the try-finally part works as expected we can extract that part and prepare for one of the next Maven releases.

Robert





On 12-11-2019 10:25:42, Stephen Connolly <[hidden email]> wrote:
On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:

> This is not just MNG-5668, but also contains several non-existing issues,
> that should be mentioned explicitly as they will have huge impact:
>
> - support before:/after: prefix for phase-binding
>
> - introduce priority
> - reduce phases (this one hasn't been implemented, but seems to be the
> reason behind before:/after:)


All detailed in the proposal on the wiki:
https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases

Reducing phases would be a big change and not before 4.x at least (maybe
5.x more realistically... at least we’d need to deprecate the phases for a
good while before removing any)


>
> I would like see separate branches for all of them, as they all have their
> own discussion.


The whole point of a PoC is the get feedback. I don’t see utility in
separate branches as they are all touching the same code.

Once we get feedback we can decide where we want to go from there.


>
> Robert
> On 11-11-2019 20:31:44, Stephen Connolly
> wrote:
> https://github.com/apache/maven/tree/mng-5668-poc is my POC implementation
> for anyone interested in trying it out.
>
> Here's a pom that builds with the PoC
>
>
> 4.0.0
> localdomain
> foo
> 1.0-SNAPSHOT
>
>
>
> maven-antrun-plugin
>
>
> 1
> before:integration-test
>
> run
>
>
>
>
>
>
>
>
> 2
> before:integration-test[1000]
>
> run
>
>
>
>
>
>
>
>
> 3
> after:integration-test
>
> run
>
>
>
>
>
>
>
>
> 4
> integration-test
>
> run
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
>
> > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > plugin-owner or pom-owner?
> >
> > I think we all recognize the issues we're trying to solve, but to me this
> > proposal is not the right solution.
> >
> > In general there are 2 issues:
> > 1. provide a mechanism that makes sure some executions are called even
> its
> > matching main phase fails.
> > 2. provide a mechanism then ensures the order of executions.
> >
> > The problem of issue 1 is described in MNG-5668, but not the final
> > solution.
> > MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> > stage 2 proposes to give the power to the *pom-owner*.
> > Both agree on the same thing: by default these post-phases should be
> > triggered even after failure of the matching main phase. This is actually
> > already expected behavior, so I don't expect real issues when
> implementing
> > this adjusted behavior.
> > To me after:integration-test is just an alias for post-integration-test,
> > both should work the same way.
> >
> > Issue 2 is a more common problem: controlling the order of executions.
> > In some cases it is pretty hard or even impossible to get the preferred
> > order. The latter happens when 2 goals of the same plugin must be
> executed
> > and a goal of another plugin are competing within the same phase.
> >
> > So let's first take a look at a phase: is there a clear definition?
> > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > lifecycle is an ordered sequence of phases involved in building a
> project".
> > "Lifecycle phases are intentionally vague, defined solely as
> > validation, testing, or deployment, and they may mean different things to
> > different projects."
> > Phases are intended to be called from the commandline, and within the pom
> > you define you can control what should happen before or during that
> phase.
> >
> > To me changing the content of the -element is a codesmell as it
> > becomes more than just a phase, and we start programming. Why do we need
> it?
> > In the end it is all about ensuring the order of plugin executions.
> > Stage3+4 proposes to give the power to the *pom-owner*,
> > whereas MPLUGIN-350[2] proposes to give this power to the *plugin-owner*.
> > IIUR Gradle does not have this issue, because their plugins are aware of
> > input and output. They ensure that if the output plugin X is the input of
> > plugin Y, than X is executed before Y.
> > And we should do the same. And this comes with benefits: we can decide if
> > executions within a project can be executed in parallel. And the pom
> stays
> > as clean as it is right now.
> >
> > In cases when there's a better ownership than the pom-owner, I would
> > prefer to choose that solution. I already notice how people (don't) build
> > up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> > know much better what they're doing.
> >
> > thanks,
> > Robert
> >
> > Some food for thoughts: consider a developer that wants to run up until
> > pre-integration-test, because he wants to bring his system in a certain
> > state so he can work with IDE to do some work.Can we say that If And Only
> > If somebody called the pre-PHASE, there's no reason to end with the
> > post-PHASE?
> >
> > [1] https://issues.apache.org/jira/browse/MNG-5668
> > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > On 26-10-2019 14:20:50, Stephen Connolly
> > wrote:
> > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> >
> > > To avoid confusion, let's call it stages.
> > >
> > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > Stage 2: before and after
> > > Stage 3: priorities (MNG-3522[2])
> > > Stage 4: transitional lifecycle
> >
> >
> > I have a prototype of stages 1-3 nearly (80%) done... just have to polish
> > up and validate the bound executions with some tests
> >
> >
> > >
> > > For both all you need to start evaluating the value of phase.
> > > For now we can assume that after:clean is just another label for
> > > post-clean and will have exactly the same effect.
> > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> that
> > > (yet). Let's start with a hardcoded list of postphases (or in case a
> goal
> > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> stage 2
> > > to make it configurable.
> > > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > > because it was bound to a phase, nor can you ask for the value of this
> > > phase. I kind of like this, plugins shouldn't care about this.
> > > However, inside Maven it will become important at which phase it is to
> > > know if there are more executions to call OR create blocks of
> executions.
> > > Now it is just a list of executions: loop and fail fast.
> > >
> > > thanks,
> > > Robert
> > >
> > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > On 25-10-2019 21:33:14, Stephen Connolly
> > > wrote:
> > > Robert,
> > >
> > > I would be fine splitting out into, pardon the pun, phases:
> > >
> > > Phase 1: before and after
> > > Phase 2: priorities
> > > Phase 3: transitional lifecycle
> > >
> > > Might have a phase 1.5 of before:* and after:* to catch the start of a
> > > lifecycle and the end of a lifecycle...
> > >
> > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > [hidden email] [mailto:
> [hidden email]
> > ]>
> > > wrote:
> > >
> > > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> > it’s
> > > fine that anyone in the community can join our Slack)
> > >
> > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > [hidden email] [mailto:
> [hidden email]
> > ]>
> > > wrote:
> > >
> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > >
> > > Thoughts?
> > > --
> > >
> > > Sent from my phone
> > > --
> > >
> > > Sent from my phone
> > > --
> > >
> > > Sent from my phone
> >
> > --
> > Sent from my phone
> >
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
On Wed 13 Nov 2019 at 19:29, Robert Scholte <[hidden email]> wrote:

> The name of the branch contains MNG-5668, but it contains much more.
> I'd likely lead to comments like "great", without being explicit saying
> which part(s).
> I am aware there's all proposals touch the same code, but can be released
> isolated from each other.
> e.g. if the enums-value are changed to "pre-" and "post-" it should work
> for the existing phases,


That would be a mistake in my opinion. We don’t know what people have
assumed about the post-integration-test phase. I think if we are addressing
finally it needs to be in a “new” phase.

Plus how do we distinguish between someone running `mvn integration-test`
and `mvn post-integration-test` for sure someone running `mvn verify`
should probably have finally behaviour in the event of the integration-test
failing... but that would require quite a bit of hacks to the build plan
whereas dynamic phases is just about decorating each phase as it is

which means we could already use it quite soon (still need to test it

> myself, though)
> I also want to provide a counter proposal, but that takes time and for me
> there are other issues more important.
>
> My biggest fear is that this will result in an All-Or-Nothing, and I like
> to prevent that. If the try-finally part works as expected we can extract
> that part and prepare for one of the next Maven releases.
>
> Robert
>
>
>
>
>
> On 12-11-2019 10:25:42, Stephen Connolly <[hidden email]>
> wrote:
> On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
>
> > This is not just MNG-5668, but also contains several non-existing issues,
> > that should be mentioned explicitly as they will have huge impact:
> >
> > - support before:/after: prefix for phase-binding
> >
> > - introduce priority
> > - reduce phases (this one hasn't been implemented, but seems to be the
> > reason behind before:/after:)
>
>
> All detailed in the proposal on the wiki:
> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>
> Reducing phases would be a big change and not before 4.x at least (maybe
> 5.x more realistically... at least we’d need to deprecate the phases for a
> good while before removing any)
>
>
> >
> > I would like see separate branches for all of them, as they all have
> their
> > own discussion.
>
>
> The whole point of a PoC is the get feedback. I don’t see utility in
> separate branches as they are all touching the same code.
>
> Once we get feedback we can decide where we want to go from there.
>
>
> >
> > Robert
> > On 11-11-2019 20:31:44, Stephen Connolly
> > wrote:
> > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> implementation
> > for anyone interested in trying it out.
> >
> > Here's a pom that builds with the PoC
> >
> >
> > 4.0.0
> > localdomain
> > foo
> > 1.0-SNAPSHOT
> >
> >
> >
> > maven-antrun-plugin
> >
> >
> > 1
> > before:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 2
> > before:integration-test[1000]
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 3
> > after:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 4
> > integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> >
> > > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > > plugin-owner or pom-owner?
> > >
> > > I think we all recognize the issues we're trying to solve, but to me
> this
> > > proposal is not the right solution.
> > >
> > > In general there are 2 issues:
> > > 1. provide a mechanism that makes sure some executions are called even
> > its
> > > matching main phase fails.
> > > 2. provide a mechanism then ensures the order of executions.
> > >
> > > The problem of issue 1 is described in MNG-5668, but not the final
> > > solution.
> > > MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> > > stage 2 proposes to give the power to the *pom-owner*.
> > > Both agree on the same thing: by default these post-phases should be
> > > triggered even after failure of the matching main phase. This is
> actually
> > > already expected behavior, so I don't expect real issues when
> > implementing
> > > this adjusted behavior.
> > > To me after:integration-test is just an alias for
> post-integration-test,
> > > both should work the same way.
> > >
> > > Issue 2 is a more common problem: controlling the order of executions.
> > > In some cases it is pretty hard or even impossible to get the preferred
> > > order. The latter happens when 2 goals of the same plugin must be
> > executed
> > > and a goal of another plugin are competing within the same phase.
> > >
> > > So let's first take a look at a phase: is there a clear definition?
> > > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > > lifecycle is an ordered sequence of phases involved in building a
> > project".
> > > "Lifecycle phases are intentionally vague, defined solely as
> > > validation, testing, or deployment, and they may mean different things
> to
> > > different projects."
> > > Phases are intended to be called from the commandline, and within the
> pom
> > > you define you can control what should happen before or during that
> > phase.
> > >
> > > To me changing the content of the -element is a codesmell as it
> > > becomes more than just a phase, and we start programming. Why do we
> need
> > it?
> > > In the end it is all about ensuring the order of plugin executions.
> > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > whereas MPLUGIN-350[2] proposes to give this power to the
> *plugin-owner*.
> > > IIUR Gradle does not have this issue, because their plugins are aware
> of
> > > input and output. They ensure that if the output plugin X is the input
> of
> > > plugin Y, than X is executed before Y.
> > > And we should do the same. And this comes with benefits: we can decide
> if
> > > executions within a project can be executed in parallel. And the pom
> > stays
> > > as clean as it is right now.
> > >
> > > In cases when there's a better ownership than the pom-owner, I would
> > > prefer to choose that solution. I already notice how people (don't)
> build
> > > up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> > > know much better what they're doing.
> > >
> > > thanks,
> > > Robert
> > >
> > > Some food for thoughts: consider a developer that wants to run up until
> > > pre-integration-test, because he wants to bring his system in a certain
> > > state so he can work with IDE to do some work.Can we say that If And
> Only
> > > If somebody called the pre-PHASE, there's no reason to end with the
> > > post-PHASE?
> > >
> > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > On 26-10-2019 14:20:50, Stephen Connolly
> > > wrote:
> > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > >
> > > > To avoid confusion, let's call it stages.
> > > >
> > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > Stage 2: before and after
> > > > Stage 3: priorities (MNG-3522[2])
> > > > Stage 4: transitional lifecycle
> > >
> > >
> > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> polish
> > > up and validate the bound executions with some tests
> > >
> > >
> > > >
> > > > For both all you need to start evaluating the value of phase.
> > > > For now we can assume that after:clean is just another label for
> > > > post-clean and will have exactly the same effect.
> > > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> > that
> > > > (yet). Let's start with a hardcoded list of postphases (or in case a
> > goal
> > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > stage 2
> > > > to make it configurable.
> > > > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > > > because it was bound to a phase, nor can you ask for the value of
> this
> > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > However, inside Maven it will become important at which phase it is
> to
> > > > know if there are more executions to call OR create blocks of
> > executions.
> > > > Now it is just a list of executions: loop and fail fast.
> > > >
> > > > thanks,
> > > > Robert
> > > >
> > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > wrote:
> > > > Robert,
> > > >
> > > > I would be fine splitting out into, pardon the pun, phases:
> > > >
> > > > Phase 1: before and after
> > > > Phase 2: priorities
> > > > Phase 3: transitional lifecycle
> > > >
> > > > Might have a phase 1.5 of before:* and after:* to catch the start of
> a
> > > > lifecycle and the end of a lifecycle...
> > > >
> > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> > > it’s
> > > > fine that anyone in the community can join our Slack)
> > > >
> > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > > >
> > > > Thoughts?
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > >
> > > --
> > > Sent from my phone
> > >
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
In reply to this post by Robert Scholte-8
On Wed 13 Nov 2019 at 19:29, Robert Scholte <[hidden email]> wrote:

> The name of the branch contains MNG-5668, but it contains much more.
> I'd likely lead to comments like "great", without being explicit saying
> which part(s).
> I am aware there's all proposals touch the same code, but can be released
> isolated from each other.
> e.g. if the enums-value are changed to "pre-" and "post-" it should work
> for the existing phases, which means we could already use it quite soon
> (still need to test it myself, though)
> I also want to provide a counter proposal, but that takes time and for me
> there are other issues more important.


How would you handle the use case that we’ve already had reported:

As a user I want to test my integration tests in my IDE by running `mvn
integration-test` so that the test environment is not torn down and I can
debug and rerun the tests until I’m ready

With the new phases, the existing pom will still work, and some user opting
into after:integration-test knows what they are getting


>
> My biggest fear is that this will result in an All-Or-Nothing, and I like
> to prevent that. If the try-finally part works as expected we can extract
> that part and prepare for one of the next Maven releases.


I’d like to understand your fear better. I’ve been playing with the PoC a
bit, and TBH it just feels right.

For sure I’d prefer a schema change to encoding in a string, but I’m also
inclined towards string encoded dependency GAVs for 5.x so that wouldn’t be
the worst if we went that way.

With pom rewriting, I think we could do a 4.1.0 model version that moved
the execution point and priority to attributes, by writing as a 4.0.0 with
the string encoded form... iow rewriting in 4.x allows us to tidy up the
schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
deployed.


>
> Robert
>
>
>
>
>
> On 12-11-2019 10:25:42, Stephen Connolly <[hidden email]>
> wrote:
> On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
>
> > This is not just MNG-5668, but also contains several non-existing issues,
> > that should be mentioned explicitly as they will have huge impact:
> >
> > - support before:/after: prefix for phase-binding
> >
> > - introduce priority
> > - reduce phases (this one hasn't been implemented, but seems to be the
> > reason behind before:/after:)
>
>
> All detailed in the proposal on the wiki:
> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>
> Reducing phases would be a big change and not before 4.x at least (maybe
> 5.x more realistically... at least we’d need to deprecate the phases for a
> good while before removing any)
>
>
> >
> > I would like see separate branches for all of them, as they all have
> their
> > own discussion.
>
>
> The whole point of a PoC is the get feedback. I don’t see utility in
> separate branches as they are all touching the same code.
>
> Once we get feedback we can decide where we want to go from there.
>
>
> >
> > Robert
> > On 11-11-2019 20:31:44, Stephen Connolly
> > wrote:
> > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> implementation
> > for anyone interested in trying it out.
> >
> > Here's a pom that builds with the PoC
> >
> >
> > 4.0.0
> > localdomain
> > foo
> > 1.0-SNAPSHOT
> >
> >
> >
> > maven-antrun-plugin
> >
> >
> > 1
> > before:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 2
> > before:integration-test[1000]
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 3
> > after:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 4
> > integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> >
> > > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > > plugin-owner or pom-owner?
> > >
> > > I think we all recognize the issues we're trying to solve, but to me
> this
> > > proposal is not the right solution.
> > >
> > > In general there are 2 issues:
> > > 1. provide a mechanism that makes sure some executions are called even
> > its
> > > matching main phase fails.
> > > 2. provide a mechanism then ensures the order of executions.
> > >
> > > The problem of issue 1 is described in MNG-5668, but not the final
> > > solution.
> > > MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> > > stage 2 proposes to give the power to the *pom-owner*.
> > > Both agree on the same thing: by default these post-phases should be
> > > triggered even after failure of the matching main phase. This is
> actually
> > > already expected behavior, so I don't expect real issues when
> > implementing
> > > this adjusted behavior.
> > > To me after:integration-test is just an alias for
> post-integration-test,
> > > both should work the same way.
> > >
> > > Issue 2 is a more common problem: controlling the order of executions.
> > > In some cases it is pretty hard or even impossible to get the preferred
> > > order. The latter happens when 2 goals of the same plugin must be
> > executed
> > > and a goal of another plugin are competing within the same phase.
> > >
> > > So let's first take a look at a phase: is there a clear definition?
> > > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > > lifecycle is an ordered sequence of phases involved in building a
> > project".
> > > "Lifecycle phases are intentionally vague, defined solely as
> > > validation, testing, or deployment, and they may mean different things
> to
> > > different projects."
> > > Phases are intended to be called from the commandline, and within the
> pom
> > > you define you can control what should happen before or during that
> > phase.
> > >
> > > To me changing the content of the -element is a codesmell as it
> > > becomes more than just a phase, and we start programming. Why do we
> need
> > it?
> > > In the end it is all about ensuring the order of plugin executions.
> > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > whereas MPLUGIN-350[2] proposes to give this power to the
> *plugin-owner*.
> > > IIUR Gradle does not have this issue, because their plugins are aware
> of
> > > input and output. They ensure that if the output plugin X is the input
> of
> > > plugin Y, than X is executed before Y.
> > > And we should do the same. And this comes with benefits: we can decide
> if
> > > executions within a project can be executed in parallel. And the pom
> > stays
> > > as clean as it is right now.
> > >
> > > In cases when there's a better ownership than the pom-owner, I would
> > > prefer to choose that solution. I already notice how people (don't)
> build
> > > up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> > > know much better what they're doing.
> > >
> > > thanks,
> > > Robert
> > >
> > > Some food for thoughts: consider a developer that wants to run up until
> > > pre-integration-test, because he wants to bring his system in a certain
> > > state so he can work with IDE to do some work.Can we say that If And
> Only
> > > If somebody called the pre-PHASE, there's no reason to end with the
> > > post-PHASE?
> > >
> > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > On 26-10-2019 14:20:50, Stephen Connolly
> > > wrote:
> > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > >
> > > > To avoid confusion, let's call it stages.
> > > >
> > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > Stage 2: before and after
> > > > Stage 3: priorities (MNG-3522[2])
> > > > Stage 4: transitional lifecycle
> > >
> > >
> > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> polish
> > > up and validate the bound executions with some tests
> > >
> > >
> > > >
> > > > For both all you need to start evaluating the value of phase.
> > > > For now we can assume that after:clean is just another label for
> > > > post-clean and will have exactly the same effect.
> > > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> > that
> > > > (yet). Let's start with a hardcoded list of postphases (or in case a
> > goal
> > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > stage 2
> > > > to make it configurable.
> > > > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > > > because it was bound to a phase, nor can you ask for the value of
> this
> > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > However, inside Maven it will become important at which phase it is
> to
> > > > know if there are more executions to call OR create blocks of
> > executions.
> > > > Now it is just a list of executions: loop and fail fast.
> > > >
> > > > thanks,
> > > > Robert
> > > >
> > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > wrote:
> > > > Robert,
> > > >
> > > > I would be fine splitting out into, pardon the pun, phases:
> > > >
> > > > Phase 1: before and after
> > > > Phase 2: priorities
> > > > Phase 3: transitional lifecycle
> > > >
> > > > Might have a phase 1.5 of before:* and after:* to catch the start of
> a
> > > > lifecycle and the end of a lifecycle...
> > > >
> > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> > > it’s
> > > > fine that anyone in the community can join our Slack)
> > > >
> > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > > >
> > > > Thoughts?
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > >
> > > --
> > > Sent from my phone
> > >
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

Robert Scholte-8
On 13-11-2019 21:46:04, Stephen Connolly <[hidden email]> wrote:
On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:

> The name of the branch contains MNG-5668, but it contains much more.
> I'd likely lead to comments like "great", without being explicit saying
> which part(s).
> I am aware there's all proposals touch the same code, but can be released
> isolated from each other.
> e.g. if the enums-value are changed to "pre-" and "post-" it should work
> for the existing phases, which means we could already use it quite soon
> (still need to test it myself, though)
> I also want to provide a counter proposal, but that takes time and for me
> there are other issues more important.


How would you handle the use case that we’ve already had reported:

As a user I want to test my integration tests in my IDE by running `mvn
integration-test` so that the test environment is not torn down and I can
debug and rerun the tests until I’m ready

Robert Scholte:
I'd say if they want to set up there environment for the integration tests, they'd be running pre-integration-test.
Next select in the IDE the test to execute. I don't see an issue here. Calling pre-integration-test implies NOT running post-integration-test.

Every time I explain people about how Maven works with phases, they are amazed it doesn't run the post-phase. I doubt we'll see issues if we switch to expected behavior.

Based on the different views, I hope to see more involvement of PMC members, because this will be a turning point that probable cannot be undone.


With the new phases, the existing pom will still work, and some user opting
into after:integration-test knows what they are getting


>
> My biggest fear is that this will result in an All-Or-Nothing, and I like
> to prevent that. If the try-finally part works as expected we can extract
> that part and prepare for one of the next Maven releases.


I’d like to understand your fear better. I’ve been playing with the PoC a
bit, and TBH it just feels right.

For sure I’d prefer a schema change to encoding in a string, but I’m also
inclined towards string encoded dependency GAVs for 5.x so that wouldn’t be
the worst if we went that way.

With pom rewriting, I think we could do a 4.1.0 model version that moved
the execution point and priority to attributes, by writing as a 4.0.0 with
the string encoded form... iow rewriting in 4.x allows us to tidy up the
schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
deployed.


>
> Robert
>
>
>
>
>
> On 12-11-2019 10:25:42, Stephen Connolly
> wrote:
> On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
>
> > This is not just MNG-5668, but also contains several non-existing issues,
> > that should be mentioned explicitly as they will have huge impact:
> >
> > - support before:/after: prefix for phase-binding
> >
> > - introduce priority
> > - reduce phases (this one hasn't been implemented, but seems to be the
> > reason behind before:/after:)
>
>
> All detailed in the proposal on the wiki:
> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>
> Reducing phases would be a big change and not before 4.x at least (maybe
> 5.x more realistically... at least we’d need to deprecate the phases for a
> good while before removing any)
>
>
> >
> > I would like see separate branches for all of them, as they all have
> their
> > own discussion.
>
>
> The whole point of a PoC is the get feedback. I don’t see utility in
> separate branches as they are all touching the same code.
>
> Once we get feedback we can decide where we want to go from there.
>
>
> >
> > Robert
> > On 11-11-2019 20:31:44, Stephen Connolly
> > wrote:
> > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> implementation
> > for anyone interested in trying it out.
> >
> > Here's a pom that builds with the PoC
> >
> >
> > 4.0.0
> > localdomain
> > foo
> > 1.0-SNAPSHOT
> >
> >
> >
> > maven-antrun-plugin
> >
> >
> > 1
> > before:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 2
> > before:integration-test[1000]
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 3
> > after:integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> > 4
> > integration-test
> >
> > run
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> >
> > > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > > plugin-owner or pom-owner?
> > >
> > > I think we all recognize the issues we're trying to solve, but to me
> this
> > > proposal is not the right solution.
> > >
> > > In general there are 2 issues:
> > > 1. provide a mechanism that makes sure some executions are called even
> > its
> > > matching main phase fails.
> > > 2. provide a mechanism then ensures the order of executions.
> > >
> > > The problem of issue 1 is described in MNG-5668, but not the final
> > > solution.
> > > MNG-5668 proposes to give this power to the *lifecycle-owner*, whereas
> > > stage 2 proposes to give the power to the *pom-owner*.
> > > Both agree on the same thing: by default these post-phases should be
> > > triggered even after failure of the matching main phase. This is
> actually
> > > already expected behavior, so I don't expect real issues when
> > implementing
> > > this adjusted behavior.
> > > To me after:integration-test is just an alias for
> post-integration-test,
> > > both should work the same way.
> > >
> > > Issue 2 is a more common problem: controlling the order of executions.
> > > In some cases it is pretty hard or even impossible to get the preferred
> > > order. The latter happens when 2 goals of the same plugin must be
> > executed
> > > and a goal of another plugin are competing within the same phase.
> > >
> > > So let's first take a look at a phase: is there a clear definition?
> > > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > > lifecycle is an ordered sequence of phases involved in building a
> > project".
> > > "Lifecycle phases are intentionally vague, defined solely as
> > > validation, testing, or deployment, and they may mean different things
> to
> > > different projects."
> > > Phases are intended to be called from the commandline, and within the
> pom
> > > you define you can control what should happen before or during that
> > phase.
> > >
> > > To me changing the content of the -element is a codesmell as it
> > > becomes more than just a phase, and we start programming. Why do we
> need
> > it?
> > > In the end it is all about ensuring the order of plugin executions.
> > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > whereas MPLUGIN-350[2] proposes to give this power to the
> *plugin-owner*.
> > > IIUR Gradle does not have this issue, because their plugins are aware
> of
> > > input and output. They ensure that if the output plugin X is the input
> of
> > > plugin Y, than X is executed before Y.
> > > And we should do the same. And this comes with benefits: we can decide
> if
> > > executions within a project can be executed in parallel. And the pom
> > stays
> > > as clean as it is right now.
> > >
> > > In cases when there's a better ownership than the pom-owner, I would
> > > prefer to choose that solution. I already notice how people (don't)
> build
> > > up their knowledge regarding poms. The lifecycle-owner and plugin-owner
> > > know much better what they're doing.
> > >
> > > thanks,
> > > Robert
> > >
> > > Some food for thoughts: consider a developer that wants to run up until
> > > pre-integration-test, because he wants to bring his system in a certain
> > > state so he can work with IDE to do some work.Can we say that If And
> Only
> > > If somebody called the pre-PHASE, there's no reason to end with the
> > > post-PHASE?
> > >
> > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > On 26-10-2019 14:20:50, Stephen Connolly
> > > wrote:
> > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > >
> > > > To avoid confusion, let's call it stages.
> > > >
> > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > Stage 2: before and after
> > > > Stage 3: priorities (MNG-3522[2])
> > > > Stage 4: transitional lifecycle
> > >
> > >
> > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> polish
> > > up and validate the bound executions with some tests
> > >
> > >
> > > >
> > > > For both all you need to start evaluating the value of phase.
> > > > For now we can assume that after:clean is just another label for
> > > > post-clean and will have exactly the same effect.
> > > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> > that
> > > > (yet). Let's start with a hardcoded list of postphases (or in case a
> > goal
> > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > stage 2
> > > > to make it configurable.
> > > > IIRC you cannot ask from inside a Mojo if is was called explicitly or
> > > > because it was bound to a phase, nor can you ask for the value of
> this
> > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > However, inside Maven it will become important at which phase it is
> to
> > > > know if there are more executions to call OR create blocks of
> > executions.
> > > > Now it is just a list of executions: loop and fail fast.
> > > >
> > > > thanks,
> > > > Robert
> > > >
> > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > wrote:
> > > > Robert,
> > > >
> > > > I would be fine splitting out into, pardon the pun, phases:
> > > >
> > > > Phase 1: before and after
> > > > Phase 2: priorities
> > > > Phase 3: transitional lifecycle
> > > >
> > > > Might have a phase 1.5 of before:* and after:* to catch the start of
> a
> > > > lifecycle and the end of a lifecycle...
> > > >
> > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra and
> > > it’s
> > > > fine that anyone in the community can join our Slack)
> > > >
> > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > [hidden email] [mailto:
> > [hidden email]
> > > ]>
> > > > wrote:
> > > >
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > > >
> > > > Thoughts?
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > > > --
> > > >
> > > > Sent from my phone
> > >
> > > --
> > > Sent from my phone
> > >
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
On Fri 15 Nov 2019 at 09:18, Robert Scholte <[hidden email]> wrote:

> On 13-11-2019 21:46:04, Stephen Connolly <[hidden email]>
> wrote:
> On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:
>
> > The name of the branch contains MNG-5668, but it contains much more.
> > I'd likely lead to comments like "great", without being explicit saying
> > which part(s).
> > I am aware there's all proposals touch the same code, but can be released
> > isolated from each other.
> > e.g. if the enums-value are changed to "pre-" and "post-" it should work
> > for the existing phases, which means we could already use it quite soon
> > (still need to test it myself, though)
> > I also want to provide a counter proposal, but that takes time and for me
> > there are other issues more important.
>
>
> How would you handle the use case that we’ve already had reported:
>
> As a user I want to test my integration tests in my IDE by running `mvn
> integration-test` so that the test environment is not torn down and I can
> debug and rerun the tests until I’m ready
>
> Robert Scholte:
> I'd say if they want to set up there environment for the integration
> tests, they'd be running pre-integration-test.
> Next select in the IDE the test to execute. I don't see an issue here.
> Calling pre-integration-test implies NOT running post-integration-test.


I disagree. I think if you run the pre- phase then you should have the
post- also run

I think we could have a differential failure mode in the pre-phases though.
Iow a pre- phase failure returns a different exit code than the actual
phase itself

>
>
> Every time I explain people about how Maven works with phases, they are
> amazed it doesn't run the post-phase. I doubt we'll see issues if we switch
> to expected behavior.
>
> Based on the different views, I hope to see more involvement of PMC
> members, because this will be a turning point that probable cannot be
> undone.
>
>
> With the new phases, the existing pom will still work, and some user opting
> into after:integration-test knows what they are getting
>
>
> >
> > My biggest fear is that this will result in an All-Or-Nothing, and I like
> > to prevent that. If the try-finally part works as expected we can extract
> > that part and prepare for one of the next Maven releases.
>
>
> I’d like to understand your fear better. I’ve been playing with the PoC a
> bit, and TBH it just feels right.
>
> For sure I’d prefer a schema change to encoding in a string, but I’m also
> inclined towards string encoded dependency GAVs for 5.x so that wouldn’t be
> the worst if we went that way.
>
> With pom rewriting, I think we could do a 4.1.0 model version that moved
> the execution point and priority to attributes, by writing as a 4.0.0 with
> the string encoded form... iow rewriting in 4.x allows us to tidy up the
> schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
> deployed.
>
>
> >
> > Robert
> >
> >
> >
> >
> >
> > On 12-11-2019 10:25:42, Stephen Connolly
> > wrote:
> > On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
> >
> > > This is not just MNG-5668, but also contains several non-existing
> issues,
> > > that should be mentioned explicitly as they will have huge impact:
> > >
> > > - support before:/after: prefix for phase-binding
> > >
> > > - introduce priority
> > > - reduce phases (this one hasn't been implemented, but seems to be the
> > > reason behind before:/after:)
> >
> >
> > All detailed in the proposal on the wiki:
> > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
> >
> > Reducing phases would be a big change and not before 4.x at least (maybe
> > 5.x more realistically... at least we’d need to deprecate the phases for
> a
> > good while before removing any)
> >
> >
> > >
> > > I would like see separate branches for all of them, as they all have
> > their
> > > own discussion.
> >
> >
> > The whole point of a PoC is the get feedback. I don’t see utility in
> > separate branches as they are all touching the same code.
> >
> > Once we get feedback we can decide where we want to go from there.
> >
> >
> > >
> > > Robert
> > > On 11-11-2019 20:31:44, Stephen Connolly
> > > wrote:
> > > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> > implementation
> > > for anyone interested in trying it out.
> > >
> > > Here's a pom that builds with the PoC
> > >
> > >
> > > 4.0.0
> > > localdomain
> > > foo
> > > 1.0-SNAPSHOT
> > >
> > >
> > >
> > > maven-antrun-plugin
> > >
> > >
> > > 1
> > > before:integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 2
> > > before:integration-test[1000]
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 3
> > > after:integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 4
> > > integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> > >
> > > > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > > > plugin-owner or pom-owner?
> > > >
> > > > I think we all recognize the issues we're trying to solve, but to me
> > this
> > > > proposal is not the right solution.
> > > >
> > > > In general there are 2 issues:
> > > > 1. provide a mechanism that makes sure some executions are called
> even
> > > its
> > > > matching main phase fails.
> > > > 2. provide a mechanism then ensures the order of executions.
> > > >
> > > > The problem of issue 1 is described in MNG-5668, but not the final
> > > > solution.
> > > > MNG-5668 proposes to give this power to the *lifecycle-owner*,
> whereas
> > > > stage 2 proposes to give the power to the *pom-owner*.
> > > > Both agree on the same thing: by default these post-phases should be
> > > > triggered even after failure of the matching main phase. This is
> > actually
> > > > already expected behavior, so I don't expect real issues when
> > > implementing
> > > > this adjusted behavior.
> > > > To me after:integration-test is just an alias for
> > post-integration-test,
> > > > both should work the same way.
> > > >
> > > > Issue 2 is a more common problem: controlling the order of
> executions.
> > > > In some cases it is pretty hard or even impossible to get the
> preferred
> > > > order. The latter happens when 2 goals of the same plugin must be
> > > executed
> > > > and a goal of another plugin are competing within the same phase.
> > > >
> > > > So let's first take a look at a phase: is there a clear definition?
> > > > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > > > lifecycle is an ordered sequence of phases involved in building a
> > > project".
> > > > "Lifecycle phases are intentionally vague, defined solely as
> > > > validation, testing, or deployment, and they may mean different
> things
> > to
> > > > different projects."
> > > > Phases are intended to be called from the commandline, and within the
> > pom
> > > > you define you can control what should happen before or during that
> > > phase.
> > > >
> > > > To me changing the content of the -element is a codesmell as it
> > > > becomes more than just a phase, and we start programming. Why do we
> > need
> > > it?
> > > > In the end it is all about ensuring the order of plugin executions.
> > > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > > whereas MPLUGIN-350[2] proposes to give this power to the
> > *plugin-owner*.
> > > > IIUR Gradle does not have this issue, because their plugins are aware
> > of
> > > > input and output. They ensure that if the output plugin X is the
> input
> > of
> > > > plugin Y, than X is executed before Y.
> > > > And we should do the same. And this comes with benefits: we can
> decide
> > if
> > > > executions within a project can be executed in parallel. And the pom
> > > stays
> > > > as clean as it is right now.
> > > >
> > > > In cases when there's a better ownership than the pom-owner, I would
> > > > prefer to choose that solution. I already notice how people (don't)
> > build
> > > > up their knowledge regarding poms. The lifecycle-owner and
> plugin-owner
> > > > know much better what they're doing.
> > > >
> > > > thanks,
> > > > Robert
> > > >
> > > > Some food for thoughts: consider a developer that wants to run up
> until
> > > > pre-integration-test, because he wants to bring his system in a
> certain
> > > > state so he can work with IDE to do some work.Can we say that If And
> > Only
> > > > If somebody called the pre-PHASE, there's no reason to end with the
> > > > post-PHASE?
> > > >
> > > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > > On 26-10-2019 14:20:50, Stephen Connolly
> > > > wrote:
> > > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > > >
> > > > > To avoid confusion, let's call it stages.
> > > > >
> > > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > > Stage 2: before and after
> > > > > Stage 3: priorities (MNG-3522[2])
> > > > > Stage 4: transitional lifecycle
> > > >
> > > >
> > > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> > polish
> > > > up and validate the bound executions with some tests
> > > >
> > > >
> > > > >
> > > > > For both all you need to start evaluating the value of phase.
> > > > > For now we can assume that after:clean is just another label for
> > > > > post-clean and will have exactly the same effect.
> > > > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> > > that
> > > > > (yet). Let's start with a hardcoded list of postphases (or in case
> a
> > > goal
> > > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > > stage 2
> > > > > to make it configurable.
> > > > > IIRC you cannot ask from inside a Mojo if is was called explicitly
> or
> > > > > because it was bound to a phase, nor can you ask for the value of
> > this
> > > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > > However, inside Maven it will become important at which phase it is
> > to
> > > > > know if there are more executions to call OR create blocks of
> > > executions.
> > > > > Now it is just a list of executions: loop and fail fast.
> > > > >
> > > > > thanks,
> > > > > Robert
> > > > >
> > > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > > wrote:
> > > > > Robert,
> > > > >
> > > > > I would be fine splitting out into, pardon the pun, phases:
> > > > >
> > > > > Phase 1: before and after
> > > > > Phase 2: priorities
> > > > > Phase 3: transitional lifecycle
> > > > >
> > > > > Might have a phase 1.5 of before:* and after:* to catch the start
> of
> > a
> > > > > lifecycle and the end of a lifecycle...
> > > > >
> > > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > > [hidden email] [mailto:
> > > [hidden email]
> > > > ]>
> > > > > wrote:
> > > > >
> > > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra
> and
> > > > it’s
> > > > > fine that anyone in the community can join our Slack)
> > > > >
> > > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > > [hidden email] [mailto:
> > > [hidden email]
> > > > ]>
> > > > > wrote:
> > > > >
> > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > > > >
> > > > > Thoughts?
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > >
> > > > --
> > > > Sent from my phone
> > > >
> > >
> > --
> > Sent from my phone
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

Robert Scholte-8
I have to admit that when trying to figure out from a Maven perspective it felt like post-X should be called with pre-X too, but that opinion has changed.
Why would anybody call pre-X? I'd say to bring the system ready to do custom X stuff, so it should stop here executing any other phases.
However, when pre-X fails, I can imagine that post-X should be called too, as Maven wasn't able to bring the system in the right state.

The problem lies in that Maven restarts the lifecycle. If only we could do something like
- run up until pre-X (pause the lifecycle execution)
- do your custom stuff
- finish with the post-X

Thinking about some kind of pause... This way at least we won't break the lifecycle and leave it clean.


On 15-11-2019 11:07:23, Stephen Connolly <[hidden email]> wrote:
On Fri 15 Nov 2019 at 09:18, Robert Scholte wrote:

> On 13-11-2019 21:46:04, Stephen Connolly
> wrote:
> On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:
>
> > The name of the branch contains MNG-5668, but it contains much more.
> > I'd likely lead to comments like "great", without being explicit saying
> > which part(s).
> > I am aware there's all proposals touch the same code, but can be released
> > isolated from each other.
> > e.g. if the enums-value are changed to "pre-" and "post-" it should work
> > for the existing phases, which means we could already use it quite soon
> > (still need to test it myself, though)
> > I also want to provide a counter proposal, but that takes time and for me
> > there are other issues more important.
>
>
> How would you handle the use case that we’ve already had reported:
>
> As a user I want to test my integration tests in my IDE by running `mvn
> integration-test` so that the test environment is not torn down and I can
> debug and rerun the tests until I’m ready
>
> Robert Scholte:
> I'd say if they want to set up there environment for the integration
> tests, they'd be running pre-integration-test.
> Next select in the IDE the test to execute. I don't see an issue here.
> Calling pre-integration-test implies NOT running post-integration-test.


I disagree. I think if you run the pre- phase then you should have the
post- also run

I think we could have a differential failure mode in the pre-phases though.
Iow a pre- phase failure returns a different exit code than the actual
phase itself

>
>
> Every time I explain people about how Maven works with phases, they are
> amazed it doesn't run the post-phase. I doubt we'll see issues if we switch
> to expected behavior.
>
> Based on the different views, I hope to see more involvement of PMC
> members, because this will be a turning point that probable cannot be
> undone.
>
>
> With the new phases, the existing pom will still work, and some user opting
> into after:integration-test knows what they are getting
>
>
> >
> > My biggest fear is that this will result in an All-Or-Nothing, and I like
> > to prevent that. If the try-finally part works as expected we can extract
> > that part and prepare for one of the next Maven releases.
>
>
> I’d like to understand your fear better. I’ve been playing with the PoC a
> bit, and TBH it just feels right.
>
> For sure I’d prefer a schema change to encoding in a string, but I’m also
> inclined towards string encoded dependency GAVs for 5.x so that wouldn’t be
> the worst if we went that way.
>
> With pom rewriting, I think we could do a 4.1.0 model version that moved
> the execution point and priority to attributes, by writing as a 4.0.0 with
> the string encoded form... iow rewriting in 4.x allows us to tidy up the
> schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
> deployed.
>
>
> >
> > Robert
> >
> >
> >
> >
> >
> > On 12-11-2019 10:25:42, Stephen Connolly
> > wrote:
> > On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
> >
> > > This is not just MNG-5668, but also contains several non-existing
> issues,
> > > that should be mentioned explicitly as they will have huge impact:
> > >
> > > - support before:/after: prefix for phase-binding
> > >
> > > - introduce priority
> > > - reduce phases (this one hasn't been implemented, but seems to be the
> > > reason behind before:/after:)
> >
> >
> > All detailed in the proposal on the wiki:
> > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
> >
> > Reducing phases would be a big change and not before 4.x at least (maybe
> > 5.x more realistically... at least we’d need to deprecate the phases for
> a
> > good while before removing any)
> >
> >
> > >
> > > I would like see separate branches for all of them, as they all have
> > their
> > > own discussion.
> >
> >
> > The whole point of a PoC is the get feedback. I don’t see utility in
> > separate branches as they are all touching the same code.
> >
> > Once we get feedback we can decide where we want to go from there.
> >
> >
> > >
> > > Robert
> > > On 11-11-2019 20:31:44, Stephen Connolly
> > > wrote:
> > > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> > implementation
> > > for anyone interested in trying it out.
> > >
> > > Here's a pom that builds with the PoC
> > >
> > >
> > > 4.0.0
> > > localdomain
> > > foo
> > > 1.0-SNAPSHOT
> > >
> > >
> > >
> > > maven-antrun-plugin
> > >
> > >
> > > 1
> > > before:integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 2
> > > before:integration-test[1000]
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 3
> > > after:integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > 4
> > > integration-test
> > >
> > > run
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> > >
> > > > TLDR: We can do better than, but who is in control? lifecycle-owner,
> > > > plugin-owner or pom-owner?
> > > >
> > > > I think we all recognize the issues we're trying to solve, but to me
> > this
> > > > proposal is not the right solution.
> > > >
> > > > In general there are 2 issues:
> > > > 1. provide a mechanism that makes sure some executions are called
> even
> > > its
> > > > matching main phase fails.
> > > > 2. provide a mechanism then ensures the order of executions.
> > > >
> > > > The problem of issue 1 is described in MNG-5668, but not the final
> > > > solution.
> > > > MNG-5668 proposes to give this power to the *lifecycle-owner*,
> whereas
> > > > stage 2 proposes to give the power to the *pom-owner*.
> > > > Both agree on the same thing: by default these post-phases should be
> > > > triggered even after failure of the matching main phase. This is
> > actually
> > > > already expected behavior, so I don't expect real issues when
> > > implementing
> > > > this adjusted behavior.
> > > > To me after:integration-test is just an alias for
> > post-integration-test,
> > > > both should work the same way.
> > > >
> > > > Issue 2 is a more common problem: controlling the order of
> executions.
> > > > In some cases it is pretty hard or even impossible to get the
> preferred
> > > > order. The latter happens when 2 goals of the same plugin must be
> > > executed
> > > > and a goal of another plugin are competing within the same phase.
> > > >
> > > > So let's first take a look at a phase: is there a clear definition?
> > > > "A phase is a step in what Maven calls a 'build lifecycle'. The build
> > > > lifecycle is an ordered sequence of phases involved in building a
> > > project".
> > > > "Lifecycle phases are intentionally vague, defined solely as
> > > > validation, testing, or deployment, and they may mean different
> things
> > to
> > > > different projects."
> > > > Phases are intended to be called from the commandline, and within the
> > pom
> > > > you define you can control what should happen before or during that
> > > phase.
> > > >
> > > > To me changing the content of the -element is a codesmell as it
> > > > becomes more than just a phase, and we start programming. Why do we
> > need
> > > it?
> > > > In the end it is all about ensuring the order of plugin executions.
> > > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > > whereas MPLUGIN-350[2] proposes to give this power to the
> > *plugin-owner*.
> > > > IIUR Gradle does not have this issue, because their plugins are aware
> > of
> > > > input and output. They ensure that if the output plugin X is the
> input
> > of
> > > > plugin Y, than X is executed before Y.
> > > > And we should do the same. And this comes with benefits: we can
> decide
> > if
> > > > executions within a project can be executed in parallel. And the pom
> > > stays
> > > > as clean as it is right now.
> > > >
> > > > In cases when there's a better ownership than the pom-owner, I would
> > > > prefer to choose that solution. I already notice how people (don't)
> > build
> > > > up their knowledge regarding poms. The lifecycle-owner and
> plugin-owner
> > > > know much better what they're doing.
> > > >
> > > > thanks,
> > > > Robert
> > > >
> > > > Some food for thoughts: consider a developer that wants to run up
> until
> > > > pre-integration-test, because he wants to bring his system in a
> certain
> > > > state so he can work with IDE to do some work.Can we say that If And
> > Only
> > > > If somebody called the pre-PHASE, there's no reason to end with the
> > > > post-PHASE?
> > > >
> > > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > > On 26-10-2019 14:20:50, Stephen Connolly
> > > > wrote:
> > > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > > >
> > > > > To avoid confusion, let's call it stages.
> > > > >
> > > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > > Stage 2: before and after
> > > > > Stage 3: priorities (MNG-3522[2])
> > > > > Stage 4: transitional lifecycle
> > > >
> > > >
> > > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> > polish
> > > > up and validate the bound executions with some tests
> > > >
> > > >
> > > > >
> > > > > For both all you need to start evaluating the value of phase.
> > > > > For now we can assume that after:clean is just another label for
> > > > > post-clean and will have exactly the same effect.
> > > > > MNG-5665 contains a proposal to change the xml, but we shouldn't do
> > > that
> > > > > (yet). Let's start with a hardcoded list of postphases (or in case
> a
> > > goal
> > > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > > stage 2
> > > > > to make it configurable.
> > > > > IIRC you cannot ask from inside a Mojo if is was called explicitly
> or
> > > > > because it was bound to a phase, nor can you ask for the value of
> > this
> > > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > > However, inside Maven it will become important at which phase it is
> > to
> > > > > know if there are more executions to call OR create blocks of
> > > executions.
> > > > > Now it is just a list of executions: loop and fail fast.
> > > > >
> > > > > thanks,
> > > > > Robert
> > > > >
> > > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > > wrote:
> > > > > Robert,
> > > > >
> > > > > I would be fine splitting out into, pardon the pun, phases:
> > > > >
> > > > > Phase 1: before and after
> > > > > Phase 2: priorities
> > > > > Phase 3: transitional lifecycle
> > > > >
> > > > > Might have a phase 1.5 of before:* and after:* to catch the start
> of
> > a
> > > > > lifecycle and the end of a lifecycle...
> > > > >
> > > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > > [hidden email] [mailto:
> > > [hidden email]
> > > > ]>
> > > > > wrote:
> > > > >
> > > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra
> and
> > > > it’s
> > > > > fine that anyone in the community can join our Slack)
> > > > >
> > > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > > [hidden email] [mailto:
> > > [hidden email]
> > > > ]>
> > > > > wrote:
> > > > >
> > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
> > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
> > > > >
> > > > > Thoughts?
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > > > --
> > > > >
> > > > > Sent from my phone
> > > >
> > > > --
> > > > Sent from my phone
> > > >
> > >
> > --
> > Sent from my phone
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
On Fri 15 Nov 2019 at 15:18, Robert Scholte <[hidden email]> wrote:

> I have to admit that when trying to figure out from a Maven perspective it
> felt like post-X should be called with pre-X too, but that opinion has
> changed.
> Why would anybody call pre-X? I'd say to bring the system ready to do
> custom X stuff, so it should stop here executing any other phases.
> However, when pre-X fails, I can imagine that post-X should be called too,
> as Maven wasn't able to bring the system in the right state.
>
> The problem lies in that Maven restarts the lifecycle. If only we could do
> something like
> - run up until pre-X (pause the lifecycle execution)
> - do your custom stuff
> - finish with the post-X
>
> Thinking about some kind of pause... This way at least we won't break the
> lifecycle and leave it clean.
>

That’s easy. Have a Maven-pause-plugin that just waits for you to press
enter. Bind it to integration-test in a profile and presto!

But that removes the need for the current explicit phases of pre- and post-

TBH I think we need to lay down the plan that we want to go towards. It
will take a while to change existing phases, in part because removing
phases is a breaking change. You can have 3rd part plugins that bind
executions to multiple phases, expecting those phases to both exist and
have specific execution behaviour.

Hence why I think we should go all the way technically, but leave the
lifecycle mostly as-is (modulo adding any new phases and flagging existing
phases as deprecated).

Half measures will only prolong to pain for users.

If instead we say: “here’s where we were, here’s where we’re going and this
is how we get there” people can incorporate that and adapt

Messing about with one phase, that’s just hacks. Adding the ability to
define phase execution guarantees... that’s where we want to go. Adding the
ability to control plugin execution order within phases... that’s where we
want to go... is the syntax where we want to go? Probably not, but it’s how
we can get there

>
>
> On 15-11-2019 11:07:23, Stephen Connolly <[hidden email]>
> wrote:
> On Fri 15 Nov 2019 at 09:18, Robert Scholte wrote:
>
> > On 13-11-2019 21:46:04, Stephen Connolly
> > wrote:
> > On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:
> >
> > > The name of the branch contains MNG-5668, but it contains much more.
> > > I'd likely lead to comments like "great", without being explicit saying
> > > which part(s).
> > > I am aware there's all proposals touch the same code, but can be
> released
> > > isolated from each other.
> > > e.g. if the enums-value are changed to "pre-" and "post-" it should
> work
> > > for the existing phases, which means we could already use it quite soon
> > > (still need to test it myself, though)
> > > I also want to provide a counter proposal, but that takes time and for
> me
> > > there are other issues more important.
> >
> >
> > How would you handle the use case that we’ve already had reported:
> >
> > As a user I want to test my integration tests in my IDE by running `mvn
> > integration-test` so that the test environment is not torn down and I can
> > debug and rerun the tests until I’m ready
> >
> > Robert Scholte:
> > I'd say if they want to set up there environment for the integration
> > tests, they'd be running pre-integration-test.
> > Next select in the IDE the test to execute. I don't see an issue here.
> > Calling pre-integration-test implies NOT running post-integration-test.
>
>
> I disagree. I think if you run the pre- phase then you should have the
> post- also run
>
> I think we could have a differential failure mode in the pre-phases though.
> Iow a pre- phase failure returns a different exit code than the actual
> phase itself
>
> >
> >
> > Every time I explain people about how Maven works with phases, they are
> > amazed it doesn't run the post-phase. I doubt we'll see issues if we
> switch
> > to expected behavior.
> >
> > Based on the different views, I hope to see more involvement of PMC
> > members, because this will be a turning point that probable cannot be
> > undone.
> >
> >
> > With the new phases, the existing pom will still work, and some user
> opting
> > into after:integration-test knows what they are getting
> >
> >
> > >
> > > My biggest fear is that this will result in an All-Or-Nothing, and I
> like
> > > to prevent that. If the try-finally part works as expected we can
> extract
> > > that part and prepare for one of the next Maven releases.
> >
> >
> > I’d like to understand your fear better. I’ve been playing with the PoC a
> > bit, and TBH it just feels right.
> >
> > For sure I’d prefer a schema change to encoding in a string, but I’m also
> > inclined towards string encoded dependency GAVs for 5.x so that wouldn’t
> be
> > the worst if we went that way.
> >
> > With pom rewriting, I think we could do a 4.1.0 model version that moved
> > the execution point and priority to attributes, by writing as a 4.0.0
> with
> > the string encoded form... iow rewriting in 4.x allows us to tidy up the
> > schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
> > deployed.
> >
> >
> > >
> > > Robert
> > >
> > >
> > >
> > >
> > >
> > > On 12-11-2019 10:25:42, Stephen Connolly
> > > wrote:
> > > On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
> > >
> > > > This is not just MNG-5668, but also contains several non-existing
> > issues,
> > > > that should be mentioned explicitly as they will have huge impact:
> > > >
> > > > - support before:/after: prefix for phase-binding
> > > >
> > > > - introduce priority
> > > > - reduce phases (this one hasn't been implemented, but seems to be
> the
> > > > reason behind before:/after:)
> > >
> > >
> > > All detailed in the proposal on the wiki:
> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
> > >
> > > Reducing phases would be a big change and not before 4.x at least
> (maybe
> > > 5.x more realistically... at least we’d need to deprecate the phases
> for
> > a
> > > good while before removing any)
> > >
> > >
> > > >
> > > > I would like see separate branches for all of them, as they all have
> > > their
> > > > own discussion.
> > >
> > >
> > > The whole point of a PoC is the get feedback. I don’t see utility in
> > > separate branches as they are all touching the same code.
> > >
> > > Once we get feedback we can decide where we want to go from there.
> > >
> > >
> > > >
> > > > Robert
> > > > On 11-11-2019 20:31:44, Stephen Connolly
> > > > wrote:
> > > > https://github.com/apache/maven/tree/mng-5668-poc is my POC
> > > implementation
> > > > for anyone interested in trying it out.
> > > >
> > > > Here's a pom that builds with the PoC
> > > >
> > > >
> > > > 4.0.0
> > > > localdomain
> > > > foo
> > > > 1.0-SNAPSHOT
> > > >
> > > >
> > > >
> > > > maven-antrun-plugin
> > > >
> > > >
> > > > 1
> > > > before:integration-test
> > > >
> > > > run
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > 2
> > > > before:integration-test[1000]
> > > >
> > > > run
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > 3
> > > > after:integration-test
> > > >
> > > > run
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > 4
> > > > integration-test
> > > >
> > > > run
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
> > > >
> > > > > TLDR: We can do better than, but who is in control?
> lifecycle-owner,
> > > > > plugin-owner or pom-owner?
> > > > >
> > > > > I think we all recognize the issues we're trying to solve, but to
> me
> > > this
> > > > > proposal is not the right solution.
> > > > >
> > > > > In general there are 2 issues:
> > > > > 1. provide a mechanism that makes sure some executions are called
> > even
> > > > its
> > > > > matching main phase fails.
> > > > > 2. provide a mechanism then ensures the order of executions.
> > > > >
> > > > > The problem of issue 1 is described in MNG-5668, but not the final
> > > > > solution.
> > > > > MNG-5668 proposes to give this power to the *lifecycle-owner*,
> > whereas
> > > > > stage 2 proposes to give the power to the *pom-owner*.
> > > > > Both agree on the same thing: by default these post-phases should
> be
> > > > > triggered even after failure of the matching main phase. This is
> > > actually
> > > > > already expected behavior, so I don't expect real issues when
> > > > implementing
> > > > > this adjusted behavior.
> > > > > To me after:integration-test is just an alias for
> > > post-integration-test,
> > > > > both should work the same way.
> > > > >
> > > > > Issue 2 is a more common problem: controlling the order of
> > executions.
> > > > > In some cases it is pretty hard or even impossible to get the
> > preferred
> > > > > order. The latter happens when 2 goals of the same plugin must be
> > > > executed
> > > > > and a goal of another plugin are competing within the same phase.
> > > > >
> > > > > So let's first take a look at a phase: is there a clear definition?
> > > > > "A phase is a step in what Maven calls a 'build lifecycle'. The
> build
> > > > > lifecycle is an ordered sequence of phases involved in building a
> > > > project".
> > > > > "Lifecycle phases are intentionally vague, defined solely as
> > > > > validation, testing, or deployment, and they may mean different
> > things
> > > to
> > > > > different projects."
> > > > > Phases are intended to be called from the commandline, and within
> the
> > > pom
> > > > > you define you can control what should happen before or during that
> > > > phase.
> > > > >
> > > > > To me changing the content of the -element is a codesmell as it
> > > > > becomes more than just a phase, and we start programming. Why do we
> > > need
> > > > it?
> > > > > In the end it is all about ensuring the order of plugin executions.
> > > > > Stage3+4 proposes to give the power to the *pom-owner*,
> > > > > whereas MPLUGIN-350[2] proposes to give this power to the
> > > *plugin-owner*.
> > > > > IIUR Gradle does not have this issue, because their plugins are
> aware
> > > of
> > > > > input and output. They ensure that if the output plugin X is the
> > input
> > > of
> > > > > plugin Y, than X is executed before Y.
> > > > > And we should do the same. And this comes with benefits: we can
> > decide
> > > if
> > > > > executions within a project can be executed in parallel. And the
> pom
> > > > stays
> > > > > as clean as it is right now.
> > > > >
> > > > > In cases when there's a better ownership than the pom-owner, I
> would
> > > > > prefer to choose that solution. I already notice how people (don't)
> > > build
> > > > > up their knowledge regarding poms. The lifecycle-owner and
> > plugin-owner
> > > > > know much better what they're doing.
> > > > >
> > > > > thanks,
> > > > > Robert
> > > > >
> > > > > Some food for thoughts: consider a developer that wants to run up
> > until
> > > > > pre-integration-test, because he wants to bring his system in a
> > certain
> > > > > state so he can work with IDE to do some work.Can we say that If
> And
> > > Only
> > > > > If somebody called the pre-PHASE, there's no reason to end with the
> > > > > post-PHASE?
> > > > >
> > > > > [1] https://issues.apache.org/jira/browse/MNG-5668
> > > > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
> > > > > On 26-10-2019 14:20:50, Stephen Connolly
> > > > > wrote:
> > > > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
> > > > >
> > > > > > To avoid confusion, let's call it stages.
> > > > > >
> > > > > > Stage 1: Always call post-bound executions (MNG-5665[1])
> > > > > > Stage 2: before and after
> > > > > > Stage 3: priorities (MNG-3522[2])
> > > > > > Stage 4: transitional lifecycle
> > > > >
> > > > >
> > > > > I have a prototype of stages 1-3 nearly (80%) done... just have to
> > > polish
> > > > > up and validate the bound executions with some tests
> > > > >
> > > > >
> > > > > >
> > > > > > For both all you need to start evaluating the value of phase.
> > > > > > For now we can assume that after:clean is just another label for
> > > > > > post-clean and will have exactly the same effect.
> > > > > > MNG-5665 contains a proposal to change the xml, but we shouldn't
> do
> > > > that
> > > > > > (yet). Let's start with a hardcoded list of postphases (or in
> case
> > a
> > > > goal
> > > > > > fails, see if a post-x phase exists). Stage 1 is to make it work,
> > > > stage 2
> > > > > > to make it configurable.
> > > > > > IIRC you cannot ask from inside a Mojo if is was called
> explicitly
> > or
> > > > > > because it was bound to a phase, nor can you ask for the value of
> > > this
> > > > > > phase. I kind of like this, plugins shouldn't care about this.
> > > > > > However, inside Maven it will become important at which phase it
> is
> > > to
> > > > > > know if there are more executions to call OR create blocks of
> > > > executions.
> > > > > > Now it is just a list of executions: loop and fail fast.
> > > > > >
> > > > > > thanks,
> > > > > > Robert
> > > > > >
> > > > > > [1] https://issues.apache.org/jira/browse/MNG-5665
> > > > > > [2] https://issues.apache.org/jira/browse/MNG-3522
> > > > > > On 25-10-2019 21:33:14, Stephen Connolly
> > > > > > wrote:
> > > > > > Robert,
> > > > > >
> > > > > > I would be fine splitting out into, pardon the pun, phases:
> > > > > >
> > > > > > Phase 1: before and after
> > > > > > Phase 2: priorities
> > > > > > Phase 3: transitional lifecycle
> > > > > >
> > > > > > Might have a phase 1.5 of before:* and after:* to catch the start
> > of
> > > a
> > > > > > lifecycle and the end of a lifecycle...
> > > > > >
> > > > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
> > > > > > [hidden email] [mailto:
> > > > [hidden email]
> > > > > ]>
> > > > > > wrote:
> > > > > >
> > > > > > Robert, Michael, Tibor, let’s continue here (though I asked Infra
> > and
> > > > > it’s
> > > > > > fine that anyone in the community can join our Slack)
> > > > > >
> > > > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
> > > > > > [hidden email] [mailto:
> > > > [hidden email]
> > > > > ]>
> > > > > > wrote:
> > > > > >
> > > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
> [
> > > > > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
> ]
> > > > > >
> > > > > > Thoughts?
> > > > > > --
> > > > > >
> > > > > > Sent from my phone
> > > > > > --
> > > > > >
> > > > > > Sent from my phone
> > > > > > --
> > > > > >
> > > > > > Sent from my phone
> > > > >
> > > > > --
> > > > > Sent from my phone
> > > > >
> > > >
> > > --
> > > Sent from my phone
> > >
> > --
> > Sent from my phone
> >
> --
> Sent from my phone
>
--
Sent from my phone
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
I have advanced the PoC a bit more by adding an experiments mechanism.

To use the dynamic phases PoC you now need to:

1. Build and install Maven on the branch
2. Add the experiments extension in .mvn/extensions.xml, e.g.

<?xml version="1.0" encoding="UTF-8"?>
<extensions xmlns="http://maven.apache.org/EXTENSIONS/1.0.0" xmlns:xsi="
http://www.w3.org/2001/XMLSchema-instance"
            xsi:schemaLocation="http://maven.apache.org/EXTENSIONS/1.0.0
http://maven.apache.org/xsd/core-extensions-1.0.0.xsd">

  <extension>
    <groupId>org.apache.maven</groupId>
    <artifactId>maven-experiments</artifactId>
    <version>3.7.0-SNAPSHOT</version>
  </extension>

</extensions>

3. Update your pom to use the new dynamic phases.

The reason for the experiments extension is to guard against assuming the
phases will work and prevent "normal" versions of Maven from producing a
bad build.

Here's a build with the extension enabled:

[INFO] Enabling experimental features of Maven 3.7.0-SNAPSHOT
[INFO] Experimental features enabled:
[INFO]   * dynamic-phases
[INFO] Scanning for projects...
[INFO]
------------------------------------------------------------------------
[INFO] Reactor Build Order:
[INFO]
[INFO] foo
 [jar]
[INFO] bar
 [jar]
[INFO] test
[pom]
[INFO]
[INFO] --------------------------< localdomain:foo
>---------------------------
[INFO] Building foo 1.0-SNAPSHOT
 [1/3]
[INFO] --------------------------------[ jar
]---------------------------------
[INFO]
[INFO] --- maven-resources-plugin:2.6:resources (default-resources) @ foo
---
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered
resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory
/Users/stephenc/tmp/test/foo/src/main/resources
[INFO]
[INFO] --- maven-compiler-plugin:3.1:compile (default-compile) @ foo ---
[INFO] No sources to compile
[INFO]
[INFO] --- maven-resources-plugin:2.6:testResources (default-testResources)
@ foo ---
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered
resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory
/Users/stephenc/tmp/test/foo/src/test/resources
[INFO]
[INFO] --- maven-compiler-plugin:3.1:testCompile (default-testCompile) @
foo ---
[INFO] No sources to compile
[INFO]
[INFO] --- maven-surefire-plugin:2.12.4:test (default-test) @ foo ---
[INFO] No tests to run.
[INFO]
[INFO] --- maven-jar-plugin:2.4:jar (default-jar) @ foo ---
[WARNING] JAR will be empty - no content was marked for inclusion!
[INFO]
[INFO] --- maven-antrun-plugin:1.3:run (2) @ foo ---
[INFO] Executing tasks
     [echo] beat you
[INFO] Executed tasks
[INFO]
[INFO] --- maven-antrun-plugin:1.3:run (1) @ foo ---
[INFO] Executing tasks
     [echo] hi
[INFO] Executed tasks
[INFO]
[INFO] --- maven-antrun-plugin:1.3:run (4) @ foo ---
[INFO] Executing tasks
[INFO]
[INFO] --- maven-antrun-plugin:1.3:run (3) @ foo ---
[INFO] Executing tasks
     [echo] bye
[INFO] Executed tasks
[INFO]
------------------------------------------------------------------------
[INFO] Reactor Summary for test 1.0-SNAPSHOT:
[INFO]
[INFO] foo ................................................ FAILURE [
 2.745 s]
[INFO] bar ................................................ SKIPPED
[INFO] test ............................................... SKIPPED
[INFO]
------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO]
------------------------------------------------------------------------
[INFO] Total time:  2.813 s
[INFO] Finished at: 2019-11-22T15:43:59Z
[INFO]
------------------------------------------------------------------------

Here's the same project with the extensions disabled

[INFO] Scanning for projects...
[INFO]
------------------------------------------------------------------------
[INFO] Reactor Build Order:
[INFO]
[INFO] foo
 [jar]
[INFO] bar
 [jar]
[INFO] test
[pom]
[INFO]
[INFO] --------------------------< localdomain:foo
>---------------------------
[INFO] Building foo 1.0-SNAPSHOT
 [1/3]
[INFO] --------------------------------[ jar
]---------------------------------
[INFO]
[INFO] --- maven-resources-plugin:2.6:resources (default-resources) @ foo
---
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered
resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory
/Users/stephenc/tmp/test/foo/src/main/resources
[INFO]
[INFO] --- maven-compiler-plugin:3.1:compile (default-compile) @ foo ---
[INFO] No sources to compile
[INFO]
[INFO] --- maven-resources-plugin:2.6:testResources (default-testResources)
@ foo ---
[WARNING] Using platform encoding (UTF-8 actually) to copy filtered
resources, i.e. build is platform dependent!
[INFO] skip non existing resourceDirectory
/Users/stephenc/tmp/test/foo/src/test/resources
[INFO]
[INFO] --- maven-compiler-plugin:3.1:testCompile (default-testCompile) @
foo ---
[INFO] No sources to compile
[INFO]
[INFO] --- maven-surefire-plugin:2.12.4:test (default-test) @ foo ---
[INFO] No tests to run.
[INFO]
[INFO] --- maven-jar-plugin:2.4:jar (default-jar) @ foo ---
[WARNING] JAR will be empty - no content was marked for inclusion!
[INFO]
[INFO] --- maven-antrun-plugin:1.3:run (4) @ foo ---
[INFO] Executing tasks
[INFO]
------------------------------------------------------------------------
[INFO] Reactor Summary for test 1.0-SNAPSHOT:
[INFO]
[INFO] foo ................................................ FAILURE [
 0.745 s]
[INFO] bar ................................................ SKIPPED
[INFO] test ............................................... SKIPPED
[INFO]
------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO]
------------------------------------------------------------------------
[INFO] Total time:  1.054 s
[INFO] Finished at: 2019-11-22T15:43:38Z
[INFO]
------------------------------------------------------------------------

Notice how the dynamic phases are completely ignored if the experiment is
not activated

And here's the foo pom.xml to show what the executions are

<project>
    <modelVersion>4.0.0</modelVersion>
    <groupId>localdomain</groupId>
    <artifactId>foo</artifactId>
    <version>1.0-SNAPSHOT</version>
    <build>
        <plugins>
            <plugin>
                <artifactId>maven-antrun-plugin</artifactId>
                <executions>
                    <execution>
                        <id>1</id>
                        <phase>before:integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <echo message="hi"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>2</id>
                        <phase>before:integration-test[1000]</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <sleep seconds="2"/>
                                <echo message="beat you"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>3</id>
                        <phase>after:integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <echo message="bye"/>
                            </tasks>
                        </configuration>
                    </execution>
                    <execution>
                        <id>4</id>
                        <phase>integration-test</phase>
                        <goals>
                            <goal>run</goal>
                        </goals>
                        <configuration>
                            <tasks>
                                <fail/>
                            </tasks>
                        </configuration>
                    </execution>
                </executions>
            </plugin>
        </plugins>
    </build>
</project>

Finally this is what you get if you have the experiment extension added but
try to build with an older version of Maven:

[WARNING] Error injecting:
org.apache.maven.feature.check.MavenExperimentEnabler
java.lang.NoClassDefFoundError:
org/apache/maven/feature/api/MavenFeatureContext
    at java.lang.Class.getDeclaredConstructors0 (Native Method)
    at java.lang.Class.privateGetDeclaredConstructors (Class.java:2671)
    at java.lang.Class.getDeclaredConstructors (Class.java:2020)
    at com.google.inject.spi.InjectionPoint.forConstructorOf
(InjectionPoint.java:245)
    at com.google.inject.internal.ConstructorBindingImpl.create
(ConstructorBindingImpl.java:115)
    at com.google.inject.internal.InjectorImpl.createUninitializedBinding
(InjectorImpl.java:706)
    at com.google.inject.internal.InjectorImpl.createJustInTimeBinding
(InjectorImpl.java:929)
    at
com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive
(InjectorImpl.java:852)
    at com.google.inject.internal.InjectorImpl.getJustInTimeBinding
(InjectorImpl.java:291)
    at com.google.inject.internal.InjectorImpl.getBindingOrThrow
(InjectorImpl.java:222)
    at com.google.inject.internal.InjectorImpl.getProviderOrThrow
(InjectorImpl.java:1040)
    at com.google.inject.internal.InjectorImpl.getProvider
(InjectorImpl.java:1071)
    at com.google.inject.internal.InjectorImpl.getProvider
(InjectorImpl.java:1034)
    at com.google.inject.internal.InjectorImpl.getInstance
(InjectorImpl.java:1086)
    at org.eclipse.sisu.space.AbstractDeferredClass.get
(AbstractDeferredClass.java:48)
    at com.google.inject.internal.ProviderInternalFactory.provision
(ProviderInternalFactory.java:85)
    at
com.google.inject.internal.InternalFactoryToInitializableAdapter.provision
(InternalFactoryToInitializableAdapter.java:57)
    at com.google.inject.internal.ProviderInternalFactory$1.call
(ProviderInternalFactory.java:66)
    at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
(ProvisionListenerStackCallback.java:112)
    at org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision
(BeanScheduler.java:230)
    at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
(ProvisionListenerStackCallback.java:120)
    at com.google.inject.internal.ProvisionListenerStackCallback.provision
(ProvisionListenerStackCallback.java:66)
    at com.google.inject.internal.ProviderInternalFactory.circularGet
(ProviderInternalFactory.java:61)
    at com.google.inject.internal.InternalFactoryToInitializableAdapter.get
(InternalFactoryToInitializableAdapter.java:47)
    at com.google.inject.internal.ProviderToInternalFactoryAdapter.get
(ProviderToInternalFactoryAdapter.java:40)
    at com.google.inject.internal.SingletonScope$1.get
(SingletonScope.java:148)
    at com.google.inject.internal.InternalFactoryToProviderAdapter.get
(InternalFactoryToProviderAdapter.java:39)
    at com.google.inject.internal.InjectorImpl$1.get
(InjectorImpl.java:1050)
    at org.eclipse.sisu.inject.LazyBeanEntry.getValue
(LazyBeanEntry.java:81)
    at org.eclipse.sisu.plexus.LazyPlexusBean.getValue
(LazyPlexusBean.java:51)
    at org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next
(EntryListAdapter.java:111)
    at java.util.AbstractCollection.addAll (AbstractCollection.java:343)
    at org.apache.maven.DefaultMaven.getLifecycleParticipants
(DefaultMaven.java:377)
    at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:206)
    at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:192)
    at org.apache.maven.DefaultMaven.execute (DefaultMaven.java:105)
    at org.apache.maven.cli.MavenCli.execute (MavenCli.java:954)
    at org.apache.maven.cli.MavenCli.doMain (MavenCli.java:288)
    at org.apache.maven.cli.MavenCli.main (MavenCli.java:192)
    at sun.reflect.NativeMethodAccessorImpl.invoke0 (Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke
(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke
(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke (Method.java:498)
    at org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced
(Launcher.java:289)
    at org.codehaus.plexus.classworlds.launcher.Launcher.launch
(Launcher.java:229)
    at org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode
(Launcher.java:415)
    at org.codehaus.plexus.classworlds.launcher.Launcher.main
(Launcher.java:356)
Caused by: java.lang.ClassNotFoundException:
org.apache.maven.feature.api.MavenFeatureContext
    at org.codehaus.plexus.classworlds.strategy.SelfFirstStrategy.loadClass
(SelfFirstStrategy.java:50)
    at
org.codehaus.plexus.classworlds.realm.ClassRealm.unsynchronizedLoadClass
(ClassRealm.java:271)
    at org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass
(ClassRealm.java:247)
    at org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass
(ClassRealm.java:239)
    at java.lang.Class.getDeclaredConstructors0 (Native Method)
    at java.lang.Class.privateGetDeclaredConstructors (Class.java:2671)
    at java.lang.Class.getDeclaredConstructors (Class.java:2020)
    at com.google.inject.spi.InjectionPoint.forConstructorOf
(InjectionPoint.java:245)
    at com.google.inject.internal.ConstructorBindingImpl.create
(ConstructorBindingImpl.java:115)
    at com.google.inject.internal.InjectorImpl.createUninitializedBinding
(InjectorImpl.java:706)
    at com.google.inject.internal.InjectorImpl.createJustInTimeBinding
(InjectorImpl.java:929)
    at
com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive
(InjectorImpl.java:852)
    at com.google.inject.internal.InjectorImpl.getJustInTimeBinding
(InjectorImpl.java:291)
    at com.google.inject.internal.InjectorImpl.getBindingOrThrow
(InjectorImpl.java:222)
    at com.google.inject.internal.InjectorImpl.getProviderOrThrow
(InjectorImpl.java:1040)
    at com.google.inject.internal.InjectorImpl.getProvider
(InjectorImpl.java:1071)
    at com.google.inject.internal.InjectorImpl.getProvider
(InjectorImpl.java:1034)
    at com.google.inject.internal.InjectorImpl.getInstance
(InjectorImpl.java:1086)
    at org.eclipse.sisu.space.AbstractDeferredClass.get
(AbstractDeferredClass.java:48)
    at com.google.inject.internal.ProviderInternalFactory.provision
(ProviderInternalFactory.java:85)
    at
com.google.inject.internal.InternalFactoryToInitializableAdapter.provision
(InternalFactoryToInitializableAdapter.java:57)
    at com.google.inject.internal.ProviderInternalFactory$1.call
(ProviderInternalFactory.java:66)
    at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
(ProvisionListenerStackCallback.java:112)
    at org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision
(BeanScheduler.java:230)
    at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
(ProvisionListenerStackCallback.java:120)
    at com.google.inject.internal.ProvisionListenerStackCallback.provision
(ProvisionListenerStackCallback.java:66)
    at com.google.inject.internal.ProviderInternalFactory.circularGet
(ProviderInternalFactory.java:61)
    at com.google.inject.internal.InternalFactoryToInitializableAdapter.get
(InternalFactoryToInitializableAdapter.java:47)
    at com.google.inject.internal.ProviderToInternalFactoryAdapter.get
(ProviderToInternalFactoryAdapter.java:40)
    at com.google.inject.internal.SingletonScope$1.get
(SingletonScope.java:148)
    at com.google.inject.internal.InternalFactoryToProviderAdapter.get
(InternalFactoryToProviderAdapter.java:39)
    at com.google.inject.internal.InjectorImpl$1.get
(InjectorImpl.java:1050)
    at org.eclipse.sisu.inject.LazyBeanEntry.getValue
(LazyBeanEntry.java:81)
    at org.eclipse.sisu.plexus.LazyPlexusBean.getValue
(LazyPlexusBean.java:51)
    at org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next
(EntryListAdapter.java:111)
    at java.util.AbstractCollection.addAll (AbstractCollection.java:343)
    at org.apache.maven.DefaultMaven.getLifecycleParticipants
(DefaultMaven.java:377)
    at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:206)
    at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:192)
    at org.apache.maven.DefaultMaven.execute (DefaultMaven.java:105)
    at org.apache.maven.cli.MavenCli.execute (MavenCli.java:954)
    at org.apache.maven.cli.MavenCli.doMain (MavenCli.java:288)
    at org.apache.maven.cli.MavenCli.main (MavenCli.java:192)
    at sun.reflect.NativeMethodAccessorImpl.invoke0 (Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke
(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke
(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke (Method.java:498)
    at org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced
(Launcher.java:289)
    at org.codehaus.plexus.classworlds.launcher.Launcher.launch
(Launcher.java:229)
    at org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode
(Launcher.java:415)
    at org.codehaus.plexus.classworlds.launcher.Launcher.main
(Launcher.java:356)
---------------------------------------------------
constituent[0]: file:/usr/local/Cellar/maven/3.5.4/libexec/conf/logging/
constituent[1]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-repository-metadata-3.5.4.jar
constituent[2]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-transport-wagon-1.1.1.jar
constituent[3]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/aopalliance-1.0.jar
constituent[4]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-provider-3.5.4.jar
constituent[5]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-utils-3.1.0.jar
constituent[6]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-interpolation-1.24.jar
constituent[7]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-artifact-3.5.4.jar
constituent[8]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-cipher-1.7.jar
constituent[9]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/guava-20.0.jar
constituent[10]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-slf4j-provider-3.5.4.jar
constituent[11]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/slf4j-api-1.7.25.jar
constituent[12]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/cdi-api-1.0.jar
constituent[13]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jcl-over-slf4j-1.7.25.jar
constituent[14]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-spi-1.1.1.jar
constituent[15]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-compat-3.5.4.jar
constituent[16]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-plugin-api-3.5.4.jar
constituent[17]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-sec-dispatcher-1.4.jar
constituent[18]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-util-1.1.1.jar
constituent[19]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-component-annotations-1.7.1.jar
constituent[20]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-settings-builder-3.5.4.jar
constituent[21]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-cli-1.4.jar
constituent[22]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-io-2.5.jar
constituent[23]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jansi-1.17.1.jar
constituent[24]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-core-3.5.4.jar
constituent[25]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-impl-1.1.1.jar
constituent[26]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/guice-4.2.0-no_aop.jar
constituent[27]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/org.eclipse.sisu.inject-0.3.3.jar
constituent[28]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-file-3.1.0.jar
constituent[29]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-builder-support-3.5.4.jar
constituent[30]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-model-3.5.4.jar
constituent[31]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-settings-3.5.4.jar
constituent[32]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-http-3.1.0-shaded.jar
constituent[33]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-api-1.1.1.jar
constituent[34]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-connector-basic-1.1.1.jar
constituent[35]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-shared-utils-3.2.1.jar
constituent[36]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/org.eclipse.sisu.plexus-0.3.3.jar
constituent[37]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-model-builder-3.5.4.jar
constituent[38]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/javax.inject-1.jar
constituent[39]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-embedder-3.5.4.jar
constituent[40]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jsr250-api-1.0.jar
constituent[41]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-lang3-3.5.jar
constituent[42]:
file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-provider-api-3.1.0.jar
---------------------------------------------------
Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/maven/feature/api/MavenFeatureContext
at java.lang.Class.getDeclaredConstructors0(Native Method)
at java.lang.Class.privateGetDeclaredConstructors(Class.java:2671)
at java.lang.Class.getDeclaredConstructors(Class.java:2020)
at
com.google.inject.spi.InjectionPoint.forConstructorOf(InjectionPoint.java:245)
at
com.google.inject.internal.ConstructorBindingImpl.create(ConstructorBindingImpl.java:115)
at
com.google.inject.internal.InjectorImpl.createUninitializedBinding(InjectorImpl.java:706)
at
com.google.inject.internal.InjectorImpl.createJustInTimeBinding(InjectorImpl.java:929)
at
com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive(InjectorImpl.java:852)
at
com.google.inject.internal.InjectorImpl.getJustInTimeBinding(InjectorImpl.java:291)
at
com.google.inject.internal.InjectorImpl.getBindingOrThrow(InjectorImpl.java:222)
at
com.google.inject.internal.InjectorImpl.getProviderOrThrow(InjectorImpl.java:1040)
at
com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1071)
at
com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1034)
at
com.google.inject.internal.InjectorImpl.getInstance(InjectorImpl.java:1086)
at
org.eclipse.sisu.space.AbstractDeferredClass.get(AbstractDeferredClass.java:48)
at
com.google.inject.internal.ProviderInternalFactory.provision(ProviderInternalFactory.java:85)
at
com.google.inject.internal.InternalFactoryToInitializableAdapter.provision(InternalFactoryToInitializableAdapter.java:57)
at
com.google.inject.internal.ProviderInternalFactory$1.call(ProviderInternalFactory.java:66)
at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision(ProvisionListenerStackCallback.java:112)
at
org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision(BeanScheduler.java:230)
at
com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision(ProvisionListenerStackCallback.java:120)
at
com.google.inject.internal.ProvisionListenerStackCallback.provision(ProvisionListenerStackCallback.java:66)
at
com.google.inject.internal.ProviderInternalFactory.circularGet(ProviderInternalFactory.java:61)
at
com.google.inject.internal.InternalFactoryToInitializableAdapter.get(InternalFactoryToInitializableAdapter.java:47)
at
com.google.inject.internal.ProviderToInternalFactoryAdapter.get(ProviderToInternalFactoryAdapter.java:40)
at com.google.inject.internal.SingletonScope$1.get(SingletonScope.java:148)
at
com.google.inject.internal.InternalFactoryToProviderAdapter.get(InternalFactoryToProviderAdapter.java:39)
at com.google.inject.internal.InjectorImpl$1.get(InjectorImpl.java:1050)
at org.eclipse.sisu.inject.LazyBeanEntry.getValue(LazyBeanEntry.java:81)
at org.eclipse.sisu.plexus.LazyPlexusBean.getValue(LazyPlexusBean.java:51)
at
org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next(EntryListAdapter.java:111)
at java.util.AbstractCollection.addAll(AbstractCollection.java:343)
at
org.apache.maven.DefaultMaven.getLifecycleParticipants(DefaultMaven.java:377)
at org.apache.maven.DefaultMaven.doExecute(DefaultMaven.java:206)
at org.apache.maven.DefaultMaven.doExecute(DefaultMaven.java:192)
at org.apache.maven.DefaultMaven.execute(DefaultMaven.java:105)
at org.apache.maven.cli.MavenCli.execute(MavenCli.java:954)
at org.apache.maven.cli.MavenCli.doMain(MavenCli.java:288)
at org.apache.maven.cli.MavenCli.main(MavenCli.java:192)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at
org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced(Launcher.java:289)
at
org.codehaus.plexus.classworlds.launcher.Launcher.launch(Launcher.java:229)
at
org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode(Launcher.java:415)
at org.codehaus.plexus.classworlds.launcher.Launcher.main(Launcher.java:356)
Caused by: java.lang.ClassNotFoundException:
org.apache.maven.feature.api.MavenFeatureContext
at
org.codehaus.plexus.classworlds.strategy.SelfFirstStrategy.loadClass(SelfFirstStrategy.java:50)
at
org.codehaus.plexus.classworlds.realm.ClassRealm.unsynchronizedLoadClass(ClassRealm.java:271)
at
org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass(ClassRealm.java:247)
at
org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass(ClassRealm.java:239)
... 47 more

I'd like to make the error message nicer, but I'll need to dig further into
Sisu. In any case it has the desired effect of preventing building a
project that uses the experiment with a version of Maven that does not
support the experimental features.

The main point of experiments is to provide a way for people to try out a
feature *that requires adapting your project to use that feature* in a safe
way that prevents users from accidentally building with a different version
of Maven. My goal would be to maybe release a
3.7.0-alpha-rfc-dynamic-phases-1 build of Maven with this experiment turned
on to gather wider feedback. Anyone using the feature would then be fully
aware that the experiment may end up different when we actually decide what
we want to do, but can then easily try it out without a big song and dance.

NOTE: the pom rewriting that Robert has scheduled for 3.7.0 is IMHO not
appropriate for this kind of experiment as it doesn't affect the actual
build behaviour. If we have implemented pom rewriting correctly, users
should not notice and shouldn't need to update their pom. Dynamic phases
does require the pom to be updated, hence why it needs a more heavy-handed
enforcement through extensions (also the enforcer plugin wouldn't guarantee
execution on all lifecycles, so if you did something in after:clean
enforcer wouldn't have run)

-Stephen

On Fri, 15 Nov 2019 at 21:16, Stephen Connolly <
[hidden email]> wrote:

>
>
> On Fri 15 Nov 2019 at 15:18, Robert Scholte <[hidden email]> wrote:
>
>> I have to admit that when trying to figure out from a Maven perspective
>> it felt like post-X should be called with pre-X too, but that opinion has
>> changed.
>> Why would anybody call pre-X? I'd say to bring the system ready to do
>> custom X stuff, so it should stop here executing any other phases.
>> However, when pre-X fails, I can imagine that post-X should be called
>> too, as Maven wasn't able to bring the system in the right state.
>>
>> The problem lies in that Maven restarts the lifecycle. If only we could
>> do something like
>> - run up until pre-X (pause the lifecycle execution)
>> - do your custom stuff
>> - finish with the post-X
>>
>> Thinking about some kind of pause... This way at least we won't break the
>> lifecycle and leave it clean.
>>
>
> That’s easy. Have a Maven-pause-plugin that just waits for you to press
> enter. Bind it to integration-test in a profile and presto!
>
> But that removes the need for the current explicit phases of pre- and post-
>
> TBH I think we need to lay down the plan that we want to go towards. It
> will take a while to change existing phases, in part because removing
> phases is a breaking change. You can have 3rd part plugins that bind
> executions to multiple phases, expecting those phases to both exist and
> have specific execution behaviour.
>
> Hence why I think we should go all the way technically, but leave the
> lifecycle mostly as-is (modulo adding any new phases and flagging existing
> phases as deprecated).
>
> Half measures will only prolong to pain for users.
>
> If instead we say: “here’s where we were, here’s where we’re going and
> this is how we get there” people can incorporate that and adapt
>
> Messing about with one phase, that’s just hacks. Adding the ability to
> define phase execution guarantees... that’s where we want to go. Adding the
> ability to control plugin execution order within phases... that’s where we
> want to go... is the syntax where we want to go? Probably not, but it’s how
> we can get there
>
>>
>>
>> On 15-11-2019 11:07:23, Stephen Connolly <[hidden email]>
>> wrote:
>> On Fri 15 Nov 2019 at 09:18, Robert Scholte wrote:
>>
>> > On 13-11-2019 21:46:04, Stephen Connolly
>> > wrote:
>> > On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:
>> >
>> > > The name of the branch contains MNG-5668, but it contains much more.
>> > > I'd likely lead to comments like "great", without being explicit
>> saying
>> > > which part(s).
>> > > I am aware there's all proposals touch the same code, but can be
>> released
>> > > isolated from each other.
>> > > e.g. if the enums-value are changed to "pre-" and "post-" it should
>> work
>> > > for the existing phases, which means we could already use it quite
>> soon
>> > > (still need to test it myself, though)
>> > > I also want to provide a counter proposal, but that takes time and
>> for me
>> > > there are other issues more important.
>> >
>> >
>> > How would you handle the use case that we’ve already had reported:
>> >
>> > As a user I want to test my integration tests in my IDE by running `mvn
>> > integration-test` so that the test environment is not torn down and I
>> can
>> > debug and rerun the tests until I’m ready
>> >
>> > Robert Scholte:
>> > I'd say if they want to set up there environment for the integration
>> > tests, they'd be running pre-integration-test.
>> > Next select in the IDE the test to execute. I don't see an issue here.
>> > Calling pre-integration-test implies NOT running post-integration-test.
>>
>>
>> I disagree. I think if you run the pre- phase then you should have the
>> post- also run
>>
>> I think we could have a differential failure mode in the pre-phases
>> though.
>> Iow a pre- phase failure returns a different exit code than the actual
>> phase itself
>>
>> >
>> >
>> > Every time I explain people about how Maven works with phases, they are
>> > amazed it doesn't run the post-phase. I doubt we'll see issues if we
>> switch
>> > to expected behavior.
>> >
>> > Based on the different views, I hope to see more involvement of PMC
>> > members, because this will be a turning point that probable cannot be
>> > undone.
>> >
>> >
>> > With the new phases, the existing pom will still work, and some user
>> opting
>> > into after:integration-test knows what they are getting
>> >
>> >
>> > >
>> > > My biggest fear is that this will result in an All-Or-Nothing, and I
>> like
>> > > to prevent that. If the try-finally part works as expected we can
>> extract
>> > > that part and prepare for one of the next Maven releases.
>> >
>> >
>> > I’d like to understand your fear better. I’ve been playing with the PoC
>> a
>> > bit, and TBH it just feels right.
>> >
>> > For sure I’d prefer a schema change to encoding in a string, but I’m
>> also
>> > inclined towards string encoded dependency GAVs for 5.x so that
>> wouldn’t be
>> > the worst if we went that way.
>> >
>> > With pom rewriting, I think we could do a 4.1.0 model version that moved
>> > the execution point and priority to attributes, by writing as a 4.0.0
>> with
>> > the string encoded form... iow rewriting in 4.x allows us to tidy up the
>> > schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that gets
>> > deployed.
>> >
>> >
>> > >
>> > > Robert
>> > >
>> > >
>> > >
>> > >
>> > >
>> > > On 12-11-2019 10:25:42, Stephen Connolly
>> > > wrote:
>> > > On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
>> > >
>> > > > This is not just MNG-5668, but also contains several non-existing
>> > issues,
>> > > > that should be mentioned explicitly as they will have huge impact:
>> > > >
>> > > > - support before:/after: prefix for phase-binding
>> > > >
>> > > > - introduce priority
>> > > > - reduce phases (this one hasn't been implemented, but seems to be
>> the
>> > > > reason behind before:/after:)
>> > >
>> > >
>> > > All detailed in the proposal on the wiki:
>> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>> > >
>> > > Reducing phases would be a big change and not before 4.x at least
>> (maybe
>> > > 5.x more realistically... at least we’d need to deprecate the phases
>> for
>> > a
>> > > good while before removing any)
>> > >
>> > >
>> > > >
>> > > > I would like see separate branches for all of them, as they all have
>> > > their
>> > > > own discussion.
>> > >
>> > >
>> > > The whole point of a PoC is the get feedback. I don’t see utility in
>> > > separate branches as they are all touching the same code.
>> > >
>> > > Once we get feedback we can decide where we want to go from there.
>> > >
>> > >
>> > > >
>> > > > Robert
>> > > > On 11-11-2019 20:31:44, Stephen Connolly
>> > > > wrote:
>> > > > https://github.com/apache/maven/tree/mng-5668-poc is my POC
>> > > implementation
>> > > > for anyone interested in trying it out.
>> > > >
>> > > > Here's a pom that builds with the PoC
>> > > >
>> > > >
>> > > > 4.0.0
>> > > > localdomain
>> > > > foo
>> > > > 1.0-SNAPSHOT
>> > > >
>> > > >
>> > > >
>> > > > maven-antrun-plugin
>> > > >
>> > > >
>> > > > 1
>> > > > before:integration-test
>> > > >
>> > > > run
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > 2
>> > > > before:integration-test[1000]
>> > > >
>> > > > run
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > 3
>> > > > after:integration-test
>> > > >
>> > > > run
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > 4
>> > > > integration-test
>> > > >
>> > > > run
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
>> > > >
>> > > > > TLDR: We can do better than, but who is in control?
>> lifecycle-owner,
>> > > > > plugin-owner or pom-owner?
>> > > > >
>> > > > > I think we all recognize the issues we're trying to solve, but to
>> me
>> > > this
>> > > > > proposal is not the right solution.
>> > > > >
>> > > > > In general there are 2 issues:
>> > > > > 1. provide a mechanism that makes sure some executions are called
>> > even
>> > > > its
>> > > > > matching main phase fails.
>> > > > > 2. provide a mechanism then ensures the order of executions.
>> > > > >
>> > > > > The problem of issue 1 is described in MNG-5668, but not the final
>> > > > > solution.
>> > > > > MNG-5668 proposes to give this power to the *lifecycle-owner*,
>> > whereas
>> > > > > stage 2 proposes to give the power to the *pom-owner*.
>> > > > > Both agree on the same thing: by default these post-phases should
>> be
>> > > > > triggered even after failure of the matching main phase. This is
>> > > actually
>> > > > > already expected behavior, so I don't expect real issues when
>> > > > implementing
>> > > > > this adjusted behavior.
>> > > > > To me after:integration-test is just an alias for
>> > > post-integration-test,
>> > > > > both should work the same way.
>> > > > >
>> > > > > Issue 2 is a more common problem: controlling the order of
>> > executions.
>> > > > > In some cases it is pretty hard or even impossible to get the
>> > preferred
>> > > > > order. The latter happens when 2 goals of the same plugin must be
>> > > > executed
>> > > > > and a goal of another plugin are competing within the same phase.
>> > > > >
>> > > > > So let's first take a look at a phase: is there a clear
>> definition?
>> > > > > "A phase is a step in what Maven calls a 'build lifecycle'. The
>> build
>> > > > > lifecycle is an ordered sequence of phases involved in building a
>> > > > project".
>> > > > > "Lifecycle phases are intentionally vague, defined solely as
>> > > > > validation, testing, or deployment, and they may mean different
>> > things
>> > > to
>> > > > > different projects."
>> > > > > Phases are intended to be called from the commandline, and within
>> the
>> > > pom
>> > > > > you define you can control what should happen before or during
>> that
>> > > > phase.
>> > > > >
>> > > > > To me changing the content of the -element is a codesmell as it
>> > > > > becomes more than just a phase, and we start programming. Why do
>> we
>> > > need
>> > > > it?
>> > > > > In the end it is all about ensuring the order of plugin
>> executions.
>> > > > > Stage3+4 proposes to give the power to the *pom-owner*,
>> > > > > whereas MPLUGIN-350[2] proposes to give this power to the
>> > > *plugin-owner*.
>> > > > > IIUR Gradle does not have this issue, because their plugins are
>> aware
>> > > of
>> > > > > input and output. They ensure that if the output plugin X is the
>> > input
>> > > of
>> > > > > plugin Y, than X is executed before Y.
>> > > > > And we should do the same. And this comes with benefits: we can
>> > decide
>> > > if
>> > > > > executions within a project can be executed in parallel. And the
>> pom
>> > > > stays
>> > > > > as clean as it is right now.
>> > > > >
>> > > > > In cases when there's a better ownership than the pom-owner, I
>> would
>> > > > > prefer to choose that solution. I already notice how people
>> (don't)
>> > > build
>> > > > > up their knowledge regarding poms. The lifecycle-owner and
>> > plugin-owner
>> > > > > know much better what they're doing.
>> > > > >
>> > > > > thanks,
>> > > > > Robert
>> > > > >
>> > > > > Some food for thoughts: consider a developer that wants to run up
>> > until
>> > > > > pre-integration-test, because he wants to bring his system in a
>> > certain
>> > > > > state so he can work with IDE to do some work.Can we say that If
>> And
>> > > Only
>> > > > > If somebody called the pre-PHASE, there's no reason to end with
>> the
>> > > > > post-PHASE?
>> > > > >
>> > > > > [1] https://issues.apache.org/jira/browse/MNG-5668
>> > > > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
>> > > > > On 26-10-2019 14:20:50, Stephen Connolly
>> > > > > wrote:
>> > > > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
>> > > > >
>> > > > > > To avoid confusion, let's call it stages.
>> > > > > >
>> > > > > > Stage 1: Always call post-bound executions (MNG-5665[1])
>> > > > > > Stage 2: before and after
>> > > > > > Stage 3: priorities (MNG-3522[2])
>> > > > > > Stage 4: transitional lifecycle
>> > > > >
>> > > > >
>> > > > > I have a prototype of stages 1-3 nearly (80%) done... just have to
>> > > polish
>> > > > > up and validate the bound executions with some tests
>> > > > >
>> > > > >
>> > > > > >
>> > > > > > For both all you need to start evaluating the value of phase.
>> > > > > > For now we can assume that after:clean is just another label for
>> > > > > > post-clean and will have exactly the same effect.
>> > > > > > MNG-5665 contains a proposal to change the xml, but we
>> shouldn't do
>> > > > that
>> > > > > > (yet). Let's start with a hardcoded list of postphases (or in
>> case
>> > a
>> > > > goal
>> > > > > > fails, see if a post-x phase exists). Stage 1 is to make it
>> work,
>> > > > stage 2
>> > > > > > to make it configurable.
>> > > > > > IIRC you cannot ask from inside a Mojo if is was called
>> explicitly
>> > or
>> > > > > > because it was bound to a phase, nor can you ask for the value
>> of
>> > > this
>> > > > > > phase. I kind of like this, plugins shouldn't care about this.
>> > > > > > However, inside Maven it will become important at which phase
>> it is
>> > > to
>> > > > > > know if there are more executions to call OR create blocks of
>> > > > executions.
>> > > > > > Now it is just a list of executions: loop and fail fast.
>> > > > > >
>> > > > > > thanks,
>> > > > > > Robert
>> > > > > >
>> > > > > > [1] https://issues.apache.org/jira/browse/MNG-5665
>> > > > > > [2] https://issues.apache.org/jira/browse/MNG-3522
>> > > > > > On 25-10-2019 21:33:14, Stephen Connolly
>> > > > > > wrote:
>> > > > > > Robert,
>> > > > > >
>> > > > > > I would be fine splitting out into, pardon the pun, phases:
>> > > > > >
>> > > > > > Phase 1: before and after
>> > > > > > Phase 2: priorities
>> > > > > > Phase 3: transitional lifecycle
>> > > > > >
>> > > > > > Might have a phase 1.5 of before:* and after:* to catch the
>> start
>> > of
>> > > a
>> > > > > > lifecycle and the end of a lifecycle...
>> > > > > >
>> > > > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
>> > > > > > [hidden email] [mailto:
>> > > > [hidden email]
>> > > > > ]>
>> > > > > > wrote:
>> > > > > >
>> > > > > > Robert, Michael, Tibor, let’s continue here (though I asked
>> Infra
>> > and
>> > > > > it’s
>> > > > > > fine that anyone in the community can join our Slack)
>> > > > > >
>> > > > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
>> > > > > > [hidden email] [mailto:
>> > > > [hidden email]
>> > > > > ]>
>> > > > > > wrote:
>> > > > > >
>> > > > > >
>> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
>> > > > > >
>> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
>> > > > > >
>> > > > > > Thoughts?
>> > > > > > --
>> > > > > >
>> > > > > > Sent from my phone
>> > > > > > --
>> > > > > >
>> > > > > > Sent from my phone
>> > > > > > --
>> > > > > >
>> > > > > > Sent from my phone
>> > > > >
>> > > > > --
>> > > > > Sent from my phone
>> > > > >
>> > > >
>> > > --
>> > > Sent from my phone
>> > >
>> > --
>> > Sent from my phone
>> >
>> --
>> Sent from my phone
>>
> --
> Sent from my phone
>
Reply | Threaded
Open this post in threaded view
|

Re: Dynamic phases proposal

stephenconnolly
Ok I figured out dynamic lookup from plexus:

$ mvn -version
Apache Maven 3.5.4 (1edded0938998edf8bf061f1ceb3cfdeccf443fe;
2018-06-17T19:33:14+01:00)
Maven home: /usr/local/Cellar/maven/3.5.4/libexec
Java version: 1.8.0_152, vendor: Oracle Corporation, runtime:
/Library/Java/JavaVirtualMachines/jdk1.8.0_152.jdk/Contents/Home/jre
Default locale: en_IE, platform encoding: UTF-8
OS name: "mac os x", version: "10.14.6", arch: "x86_64", family: "mac"
$ mvn validate
[ERROR] The project uses experimental features that require exactly Maven
3.7.0-SNAPSHOT -> [Help 1]
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e
switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions,
please read the following articles:
[ERROR] [Help 1]
http://cwiki.apache.org/confluence/display/MAVEN/MavenExecutionException

Much nicer!



On Fri, 22 Nov 2019 at 16:12, Stephen Connolly <
[hidden email]> wrote:

> I have advanced the PoC a bit more by adding an experiments mechanism.
>
> To use the dynamic phases PoC you now need to:
>
> 1. Build and install Maven on the branch
> 2. Add the experiments extension in .mvn/extensions.xml, e.g.
>
> <?xml version="1.0" encoding="UTF-8"?>
> <extensions xmlns="http://maven.apache.org/EXTENSIONS/1.0.0" xmlns:xsi="
> http://www.w3.org/2001/XMLSchema-instance"
>             xsi:schemaLocation="http://maven.apache.org/EXTENSIONS/1.0.0
> http://maven.apache.org/xsd/core-extensions-1.0.0.xsd">
>
>   <extension>
>     <groupId>org.apache.maven</groupId>
>     <artifactId>maven-experiments</artifactId>
>     <version>3.7.0-SNAPSHOT</version>
>   </extension>
>
> </extensions>
>
> 3. Update your pom to use the new dynamic phases.
>
> The reason for the experiments extension is to guard against assuming the
> phases will work and prevent "normal" versions of Maven from producing a
> bad build.
>
> Here's a build with the extension enabled:
>
> [INFO] Enabling experimental features of Maven 3.7.0-SNAPSHOT
> [INFO] Experimental features enabled:
> [INFO]   * dynamic-phases
> [INFO] Scanning for projects...
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Reactor Build Order:
> [INFO]
> [INFO] foo
>  [jar]
> [INFO] bar
>  [jar]
> [INFO] test
> [pom]
> [INFO]
> [INFO] --------------------------< localdomain:foo
> >---------------------------
> [INFO] Building foo 1.0-SNAPSHOT
>  [1/3]
> [INFO] --------------------------------[ jar
> ]---------------------------------
> [INFO]
> [INFO] --- maven-resources-plugin:2.6:resources (default-resources) @ foo
> ---
> [WARNING] Using platform encoding (UTF-8 actually) to copy filtered
> resources, i.e. build is platform dependent!
> [INFO] skip non existing resourceDirectory
> /Users/stephenc/tmp/test/foo/src/main/resources
> [INFO]
> [INFO] --- maven-compiler-plugin:3.1:compile (default-compile) @ foo ---
> [INFO] No sources to compile
> [INFO]
> [INFO] --- maven-resources-plugin:2.6:testResources
> (default-testResources) @ foo ---
> [WARNING] Using platform encoding (UTF-8 actually) to copy filtered
> resources, i.e. build is platform dependent!
> [INFO] skip non existing resourceDirectory
> /Users/stephenc/tmp/test/foo/src/test/resources
> [INFO]
> [INFO] --- maven-compiler-plugin:3.1:testCompile (default-testCompile) @
> foo ---
> [INFO] No sources to compile
> [INFO]
> [INFO] --- maven-surefire-plugin:2.12.4:test (default-test) @ foo ---
> [INFO] No tests to run.
> [INFO]
> [INFO] --- maven-jar-plugin:2.4:jar (default-jar) @ foo ---
> [WARNING] JAR will be empty - no content was marked for inclusion!
> [INFO]
> [INFO] --- maven-antrun-plugin:1.3:run (2) @ foo ---
> [INFO] Executing tasks
>      [echo] beat you
> [INFO] Executed tasks
> [INFO]
> [INFO] --- maven-antrun-plugin:1.3:run (1) @ foo ---
> [INFO] Executing tasks
>      [echo] hi
> [INFO] Executed tasks
> [INFO]
> [INFO] --- maven-antrun-plugin:1.3:run (4) @ foo ---
> [INFO] Executing tasks
> [INFO]
> [INFO] --- maven-antrun-plugin:1.3:run (3) @ foo ---
> [INFO] Executing tasks
>      [echo] bye
> [INFO] Executed tasks
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Reactor Summary for test 1.0-SNAPSHOT:
> [INFO]
> [INFO] foo ................................................ FAILURE [
>  2.745 s]
> [INFO] bar ................................................ SKIPPED
> [INFO] test ............................................... SKIPPED
> [INFO]
> ------------------------------------------------------------------------
> [INFO] BUILD FAILURE
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Total time:  2.813 s
> [INFO] Finished at: 2019-11-22T15:43:59Z
> [INFO]
> ------------------------------------------------------------------------
>
> Here's the same project with the extensions disabled
>
> [INFO] Scanning for projects...
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Reactor Build Order:
> [INFO]
> [INFO] foo
>  [jar]
> [INFO] bar
>  [jar]
> [INFO] test
> [pom]
> [INFO]
> [INFO] --------------------------< localdomain:foo
> >---------------------------
> [INFO] Building foo 1.0-SNAPSHOT
>  [1/3]
> [INFO] --------------------------------[ jar
> ]---------------------------------
> [INFO]
> [INFO] --- maven-resources-plugin:2.6:resources (default-resources) @ foo
> ---
> [WARNING] Using platform encoding (UTF-8 actually) to copy filtered
> resources, i.e. build is platform dependent!
> [INFO] skip non existing resourceDirectory
> /Users/stephenc/tmp/test/foo/src/main/resources
> [INFO]
> [INFO] --- maven-compiler-plugin:3.1:compile (default-compile) @ foo ---
> [INFO] No sources to compile
> [INFO]
> [INFO] --- maven-resources-plugin:2.6:testResources
> (default-testResources) @ foo ---
> [WARNING] Using platform encoding (UTF-8 actually) to copy filtered
> resources, i.e. build is platform dependent!
> [INFO] skip non existing resourceDirectory
> /Users/stephenc/tmp/test/foo/src/test/resources
> [INFO]
> [INFO] --- maven-compiler-plugin:3.1:testCompile (default-testCompile) @
> foo ---
> [INFO] No sources to compile
> [INFO]
> [INFO] --- maven-surefire-plugin:2.12.4:test (default-test) @ foo ---
> [INFO] No tests to run.
> [INFO]
> [INFO] --- maven-jar-plugin:2.4:jar (default-jar) @ foo ---
> [WARNING] JAR will be empty - no content was marked for inclusion!
> [INFO]
> [INFO] --- maven-antrun-plugin:1.3:run (4) @ foo ---
> [INFO] Executing tasks
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Reactor Summary for test 1.0-SNAPSHOT:
> [INFO]
> [INFO] foo ................................................ FAILURE [
>  0.745 s]
> [INFO] bar ................................................ SKIPPED
> [INFO] test ............................................... SKIPPED
> [INFO]
> ------------------------------------------------------------------------
> [INFO] BUILD FAILURE
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Total time:  1.054 s
> [INFO] Finished at: 2019-11-22T15:43:38Z
> [INFO]
> ------------------------------------------------------------------------
>
> Notice how the dynamic phases are completely ignored if the experiment is
> not activated
>
> And here's the foo pom.xml to show what the executions are
>
> <project>
>     <modelVersion>4.0.0</modelVersion>
>     <groupId>localdomain</groupId>
>     <artifactId>foo</artifactId>
>     <version>1.0-SNAPSHOT</version>
>     <build>
>         <plugins>
>             <plugin>
>                 <artifactId>maven-antrun-plugin</artifactId>
>                 <executions>
>                     <execution>
>                         <id>1</id>
>                         <phase>before:integration-test</phase>
>                         <goals>
>                             <goal>run</goal>
>                         </goals>
>                         <configuration>
>                             <tasks>
>                                 <echo message="hi"/>
>                             </tasks>
>                         </configuration>
>                     </execution>
>                     <execution>
>                         <id>2</id>
>                         <phase>before:integration-test[1000]</phase>
>                         <goals>
>                             <goal>run</goal>
>                         </goals>
>                         <configuration>
>                             <tasks>
>                                 <sleep seconds="2"/>
>                                 <echo message="beat you"/>
>                             </tasks>
>                         </configuration>
>                     </execution>
>                     <execution>
>                         <id>3</id>
>                         <phase>after:integration-test</phase>
>                         <goals>
>                             <goal>run</goal>
>                         </goals>
>                         <configuration>
>                             <tasks>
>                                 <echo message="bye"/>
>                             </tasks>
>                         </configuration>
>                     </execution>
>                     <execution>
>                         <id>4</id>
>                         <phase>integration-test</phase>
>                         <goals>
>                             <goal>run</goal>
>                         </goals>
>                         <configuration>
>                             <tasks>
>                                 <fail/>
>                             </tasks>
>                         </configuration>
>                     </execution>
>                 </executions>
>             </plugin>
>         </plugins>
>     </build>
> </project>
>
> Finally this is what you get if you have the experiment extension added
> but try to build with an older version of Maven:
>
> [WARNING] Error injecting:
> org.apache.maven.feature.check.MavenExperimentEnabler
> java.lang.NoClassDefFoundError:
> org/apache/maven/feature/api/MavenFeatureContext
>     at java.lang.Class.getDeclaredConstructors0 (Native Method)
>     at java.lang.Class.privateGetDeclaredConstructors (Class.java:2671)
>     at java.lang.Class.getDeclaredConstructors (Class.java:2020)
>     at com.google.inject.spi.InjectionPoint.forConstructorOf
> (InjectionPoint.java:245)
>     at com.google.inject.internal.ConstructorBindingImpl.create
> (ConstructorBindingImpl.java:115)
>     at com.google.inject.internal.InjectorImpl.createUninitializedBinding
> (InjectorImpl.java:706)
>     at com.google.inject.internal.InjectorImpl.createJustInTimeBinding
> (InjectorImpl.java:929)
>     at
> com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive
> (InjectorImpl.java:852)
>     at com.google.inject.internal.InjectorImpl.getJustInTimeBinding
> (InjectorImpl.java:291)
>     at com.google.inject.internal.InjectorImpl.getBindingOrThrow
> (InjectorImpl.java:222)
>     at com.google.inject.internal.InjectorImpl.getProviderOrThrow
> (InjectorImpl.java:1040)
>     at com.google.inject.internal.InjectorImpl.getProvider
> (InjectorImpl.java:1071)
>     at com.google.inject.internal.InjectorImpl.getProvider
> (InjectorImpl.java:1034)
>     at com.google.inject.internal.InjectorImpl.getInstance
> (InjectorImpl.java:1086)
>     at org.eclipse.sisu.space.AbstractDeferredClass.get
> (AbstractDeferredClass.java:48)
>     at com.google.inject.internal.ProviderInternalFactory.provision
> (ProviderInternalFactory.java:85)
>     at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.provision
> (InternalFactoryToInitializableAdapter.java:57)
>     at com.google.inject.internal.ProviderInternalFactory$1.call
> (ProviderInternalFactory.java:66)
>     at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
> (ProvisionListenerStackCallback.java:112)
>     at org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision
> (BeanScheduler.java:230)
>     at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
> (ProvisionListenerStackCallback.java:120)
>     at com.google.inject.internal.ProvisionListenerStackCallback.provision
> (ProvisionListenerStackCallback.java:66)
>     at com.google.inject.internal.ProviderInternalFactory.circularGet
> (ProviderInternalFactory.java:61)
>     at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.get
> (InternalFactoryToInitializableAdapter.java:47)
>     at com.google.inject.internal.ProviderToInternalFactoryAdapter.get
> (ProviderToInternalFactoryAdapter.java:40)
>     at com.google.inject.internal.SingletonScope$1.get
> (SingletonScope.java:148)
>     at com.google.inject.internal.InternalFactoryToProviderAdapter.get
> (InternalFactoryToProviderAdapter.java:39)
>     at com.google.inject.internal.InjectorImpl$1.get
> (InjectorImpl.java:1050)
>     at org.eclipse.sisu.inject.LazyBeanEntry.getValue
> (LazyBeanEntry.java:81)
>     at org.eclipse.sisu.plexus.LazyPlexusBean.getValue
> (LazyPlexusBean.java:51)
>     at org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next
> (EntryListAdapter.java:111)
>     at java.util.AbstractCollection.addAll (AbstractCollection.java:343)
>     at org.apache.maven.DefaultMaven.getLifecycleParticipants
> (DefaultMaven.java:377)
>     at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:206)
>     at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:192)
>     at org.apache.maven.DefaultMaven.execute (DefaultMaven.java:105)
>     at org.apache.maven.cli.MavenCli.execute (MavenCli.java:954)
>     at org.apache.maven.cli.MavenCli.doMain (MavenCli.java:288)
>     at org.apache.maven.cli.MavenCli.main (MavenCli.java:192)
>     at sun.reflect.NativeMethodAccessorImpl.invoke0 (Native Method)
>     at sun.reflect.NativeMethodAccessorImpl.invoke
> (NativeMethodAccessorImpl.java:62)
>     at sun.reflect.DelegatingMethodAccessorImpl.invoke
> (DelegatingMethodAccessorImpl.java:43)
>     at java.lang.reflect.Method.invoke (Method.java:498)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced
> (Launcher.java:289)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.launch
> (Launcher.java:229)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode
> (Launcher.java:415)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.main
> (Launcher.java:356)
> Caused by: java.lang.ClassNotFoundException:
> org.apache.maven.feature.api.MavenFeatureContext
>     at
> org.codehaus.plexus.classworlds.strategy.SelfFirstStrategy.loadClass
> (SelfFirstStrategy.java:50)
>     at
> org.codehaus.plexus.classworlds.realm.ClassRealm.unsynchronizedLoadClass
> (ClassRealm.java:271)
>     at org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass
> (ClassRealm.java:247)
>     at org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass
> (ClassRealm.java:239)
>     at java.lang.Class.getDeclaredConstructors0 (Native Method)
>     at java.lang.Class.privateGetDeclaredConstructors (Class.java:2671)
>     at java.lang.Class.getDeclaredConstructors (Class.java:2020)
>     at com.google.inject.spi.InjectionPoint.forConstructorOf
> (InjectionPoint.java:245)
>     at com.google.inject.internal.ConstructorBindingImpl.create
> (ConstructorBindingImpl.java:115)
>     at com.google.inject.internal.InjectorImpl.createUninitializedBinding
> (InjectorImpl.java:706)
>     at com.google.inject.internal.InjectorImpl.createJustInTimeBinding
> (InjectorImpl.java:929)
>     at
> com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive
> (InjectorImpl.java:852)
>     at com.google.inject.internal.InjectorImpl.getJustInTimeBinding
> (InjectorImpl.java:291)
>     at com.google.inject.internal.InjectorImpl.getBindingOrThrow
> (InjectorImpl.java:222)
>     at com.google.inject.internal.InjectorImpl.getProviderOrThrow
> (InjectorImpl.java:1040)
>     at com.google.inject.internal.InjectorImpl.getProvider
> (InjectorImpl.java:1071)
>     at com.google.inject.internal.InjectorImpl.getProvider
> (InjectorImpl.java:1034)
>     at com.google.inject.internal.InjectorImpl.getInstance
> (InjectorImpl.java:1086)
>     at org.eclipse.sisu.space.AbstractDeferredClass.get
> (AbstractDeferredClass.java:48)
>     at com.google.inject.internal.ProviderInternalFactory.provision
> (ProviderInternalFactory.java:85)
>     at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.provision
> (InternalFactoryToInitializableAdapter.java:57)
>     at com.google.inject.internal.ProviderInternalFactory$1.call
> (ProviderInternalFactory.java:66)
>     at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
> (ProvisionListenerStackCallback.java:112)
>     at org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision
> (BeanScheduler.java:230)
>     at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision
> (ProvisionListenerStackCallback.java:120)
>     at com.google.inject.internal.ProvisionListenerStackCallback.provision
> (ProvisionListenerStackCallback.java:66)
>     at com.google.inject.internal.ProviderInternalFactory.circularGet
> (ProviderInternalFactory.java:61)
>     at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.get
> (InternalFactoryToInitializableAdapter.java:47)
>     at com.google.inject.internal.ProviderToInternalFactoryAdapter.get
> (ProviderToInternalFactoryAdapter.java:40)
>     at com.google.inject.internal.SingletonScope$1.get
> (SingletonScope.java:148)
>     at com.google.inject.internal.InternalFactoryToProviderAdapter.get
> (InternalFactoryToProviderAdapter.java:39)
>     at com.google.inject.internal.InjectorImpl$1.get
> (InjectorImpl.java:1050)
>     at org.eclipse.sisu.inject.LazyBeanEntry.getValue
> (LazyBeanEntry.java:81)
>     at org.eclipse.sisu.plexus.LazyPlexusBean.getValue
> (LazyPlexusBean.java:51)
>     at org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next
> (EntryListAdapter.java:111)
>     at java.util.AbstractCollection.addAll (AbstractCollection.java:343)
>     at org.apache.maven.DefaultMaven.getLifecycleParticipants
> (DefaultMaven.java:377)
>     at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:206)
>     at org.apache.maven.DefaultMaven.doExecute (DefaultMaven.java:192)
>     at org.apache.maven.DefaultMaven.execute (DefaultMaven.java:105)
>     at org.apache.maven.cli.MavenCli.execute (MavenCli.java:954)
>     at org.apache.maven.cli.MavenCli.doMain (MavenCli.java:288)
>     at org.apache.maven.cli.MavenCli.main (MavenCli.java:192)
>     at sun.reflect.NativeMethodAccessorImpl.invoke0 (Native Method)
>     at sun.reflect.NativeMethodAccessorImpl.invoke
> (NativeMethodAccessorImpl.java:62)
>     at sun.reflect.DelegatingMethodAccessorImpl.invoke
> (DelegatingMethodAccessorImpl.java:43)
>     at java.lang.reflect.Method.invoke (Method.java:498)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced
> (Launcher.java:289)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.launch
> (Launcher.java:229)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode
> (Launcher.java:415)
>     at org.codehaus.plexus.classworlds.launcher.Launcher.main
> (Launcher.java:356)
> ---------------------------------------------------
> constituent[0]: file:/usr/local/Cellar/maven/3.5.4/libexec/conf/logging/
> constituent[1]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-repository-metadata-3.5.4.jar
> constituent[2]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-transport-wagon-1.1.1.jar
> constituent[3]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/aopalliance-1.0.jar
> constituent[4]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-provider-3.5.4.jar
> constituent[5]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-utils-3.1.0.jar
> constituent[6]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-interpolation-1.24.jar
> constituent[7]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-artifact-3.5.4.jar
> constituent[8]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-cipher-1.7.jar
> constituent[9]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/guava-20.0.jar
> constituent[10]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-slf4j-provider-3.5.4.jar
> constituent[11]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/slf4j-api-1.7.25.jar
> constituent[12]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/cdi-api-1.0.jar
> constituent[13]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jcl-over-slf4j-1.7.25.jar
> constituent[14]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-spi-1.1.1.jar
> constituent[15]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-compat-3.5.4.jar
> constituent[16]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-plugin-api-3.5.4.jar
> constituent[17]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-sec-dispatcher-1.4.jar
> constituent[18]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-util-1.1.1.jar
> constituent[19]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/plexus-component-annotations-1.7.1.jar
> constituent[20]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-settings-builder-3.5.4.jar
> constituent[21]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-cli-1.4.jar
> constituent[22]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-io-2.5.jar
> constituent[23]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jansi-1.17.1.jar
> constituent[24]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-core-3.5.4.jar
> constituent[25]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-impl-1.1.1.jar
> constituent[26]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/guice-4.2.0-no_aop.jar
> constituent[27]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/org.eclipse.sisu.inject-0.3.3.jar
> constituent[28]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-file-3.1.0.jar
> constituent[29]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-builder-support-3.5.4.jar
> constituent[30]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-model-3.5.4.jar
> constituent[31]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-settings-3.5.4.jar
> constituent[32]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-http-3.1.0-shaded.jar
> constituent[33]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-api-1.1.1.jar
> constituent[34]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-resolver-connector-basic-1.1.1.jar
> constituent[35]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-shared-utils-3.2.1.jar
> constituent[36]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/org.eclipse.sisu.plexus-0.3.3.jar
> constituent[37]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-model-builder-3.5.4.jar
> constituent[38]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/javax.inject-1.jar
> constituent[39]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/maven-embedder-3.5.4.jar
> constituent[40]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/jsr250-api-1.0.jar
> constituent[41]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/commons-lang3-3.5.jar
> constituent[42]:
> file:/usr/local/Cellar/maven/3.5.4/libexec/lib/wagon-provider-api-3.1.0.jar
> ---------------------------------------------------
> Exception in thread "main" java.lang.NoClassDefFoundError:
> org/apache/maven/feature/api/MavenFeatureContext
> at java.lang.Class.getDeclaredConstructors0(Native Method)
> at java.lang.Class.privateGetDeclaredConstructors(Class.java:2671)
> at java.lang.Class.getDeclaredConstructors(Class.java:2020)
> at
> com.google.inject.spi.InjectionPoint.forConstructorOf(InjectionPoint.java:245)
> at
> com.google.inject.internal.ConstructorBindingImpl.create(ConstructorBindingImpl.java:115)
> at
> com.google.inject.internal.InjectorImpl.createUninitializedBinding(InjectorImpl.java:706)
> at
> com.google.inject.internal.InjectorImpl.createJustInTimeBinding(InjectorImpl.java:929)
> at
> com.google.inject.internal.InjectorImpl.createJustInTimeBindingRecursive(InjectorImpl.java:852)
> at
> com.google.inject.internal.InjectorImpl.getJustInTimeBinding(InjectorImpl.java:291)
> at
> com.google.inject.internal.InjectorImpl.getBindingOrThrow(InjectorImpl.java:222)
> at
> com.google.inject.internal.InjectorImpl.getProviderOrThrow(InjectorImpl.java:1040)
> at
> com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1071)
> at
> com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1034)
> at
> com.google.inject.internal.InjectorImpl.getInstance(InjectorImpl.java:1086)
> at
> org.eclipse.sisu.space.AbstractDeferredClass.get(AbstractDeferredClass.java:48)
> at
> com.google.inject.internal.ProviderInternalFactory.provision(ProviderInternalFactory.java:85)
> at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.provision(InternalFactoryToInitializableAdapter.java:57)
> at
> com.google.inject.internal.ProviderInternalFactory$1.call(ProviderInternalFactory.java:66)
> at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision(ProvisionListenerStackCallback.java:112)
> at
> org.eclipse.sisu.bean.BeanScheduler$CycleActivator.onProvision(BeanScheduler.java:230)
> at
> com.google.inject.internal.ProvisionListenerStackCallback$Provision.provision(ProvisionListenerStackCallback.java:120)
> at
> com.google.inject.internal.ProvisionListenerStackCallback.provision(ProvisionListenerStackCallback.java:66)
> at
> com.google.inject.internal.ProviderInternalFactory.circularGet(ProviderInternalFactory.java:61)
> at
> com.google.inject.internal.InternalFactoryToInitializableAdapter.get(InternalFactoryToInitializableAdapter.java:47)
> at
> com.google.inject.internal.ProviderToInternalFactoryAdapter.get(ProviderToInternalFactoryAdapter.java:40)
> at com.google.inject.internal.SingletonScope$1.get(SingletonScope.java:148)
> at
> com.google.inject.internal.InternalFactoryToProviderAdapter.get(InternalFactoryToProviderAdapter.java:39)
> at com.google.inject.internal.InjectorImpl$1.get(InjectorImpl.java:1050)
> at org.eclipse.sisu.inject.LazyBeanEntry.getValue(LazyBeanEntry.java:81)
> at org.eclipse.sisu.plexus.LazyPlexusBean.getValue(LazyPlexusBean.java:51)
> at
> org.eclipse.sisu.wire.EntryListAdapter$ValueIterator.next(EntryListAdapter.java:111)
> at java.util.AbstractCollection.addAll(AbstractCollection.java:343)
> at
> org.apache.maven.DefaultMaven.getLifecycleParticipants(DefaultMaven.java:377)
> at org.apache.maven.DefaultMaven.doExecute(DefaultMaven.java:206)
> at org.apache.maven.DefaultMaven.doExecute(DefaultMaven.java:192)
> at org.apache.maven.DefaultMaven.execute(DefaultMaven.java:105)
> at org.apache.maven.cli.MavenCli.execute(MavenCli.java:954)
> at org.apache.maven.cli.MavenCli.doMain(MavenCli.java:288)
> at org.apache.maven.cli.MavenCli.main(MavenCli.java:192)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:498)
> at
> org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced(Launcher.java:289)
> at
> org.codehaus.plexus.classworlds.launcher.Launcher.launch(Launcher.java:229)
> at
> org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode(Launcher.java:415)
> at
> org.codehaus.plexus.classworlds.launcher.Launcher.main(Launcher.java:356)
> Caused by: java.lang.ClassNotFoundException:
> org.apache.maven.feature.api.MavenFeatureContext
> at
> org.codehaus.plexus.classworlds.strategy.SelfFirstStrategy.loadClass(SelfFirstStrategy.java:50)
> at
> org.codehaus.plexus.classworlds.realm.ClassRealm.unsynchronizedLoadClass(ClassRealm.java:271)
> at
> org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass(ClassRealm.java:247)
> at
> org.codehaus.plexus.classworlds.realm.ClassRealm.loadClass(ClassRealm.java:239)
> ... 47 more
>
> I'd like to make the error message nicer, but I'll need to dig further
> into Sisu. In any case it has the desired effect of preventing building a
> project that uses the experiment with a version of Maven that does not
> support the experimental features.
>
> The main point of experiments is to provide a way for people to try out a
> feature *that requires adapting your project to use that feature* in a safe
> way that prevents users from accidentally building with a different version
> of Maven. My goal would be to maybe release a
> 3.7.0-alpha-rfc-dynamic-phases-1 build of Maven with this experiment turned
> on to gather wider feedback. Anyone using the feature would then be fully
> aware that the experiment may end up different when we actually decide what
> we want to do, but can then easily try it out without a big song and dance.
>
> NOTE: the pom rewriting that Robert has scheduled for 3.7.0 is IMHO not
> appropriate for this kind of experiment as it doesn't affect the actual
> build behaviour. If we have implemented pom rewriting correctly, users
> should not notice and shouldn't need to update their pom. Dynamic phases
> does require the pom to be updated, hence why it needs a more heavy-handed
> enforcement through extensions (also the enforcer plugin wouldn't guarantee
> execution on all lifecycles, so if you did something in after:clean
> enforcer wouldn't have run)
>
> -Stephen
>
> On Fri, 15 Nov 2019 at 21:16, Stephen Connolly <
> [hidden email]> wrote:
>
>>
>>
>> On Fri 15 Nov 2019 at 15:18, Robert Scholte <[hidden email]> wrote:
>>
>>> I have to admit that when trying to figure out from a Maven perspective
>>> it felt like post-X should be called with pre-X too, but that opinion has
>>> changed.
>>> Why would anybody call pre-X? I'd say to bring the system ready to do
>>> custom X stuff, so it should stop here executing any other phases.
>>> However, when pre-X fails, I can imagine that post-X should be called
>>> too, as Maven wasn't able to bring the system in the right state.
>>>
>>> The problem lies in that Maven restarts the lifecycle. If only we could
>>> do something like
>>> - run up until pre-X (pause the lifecycle execution)
>>> - do your custom stuff
>>> - finish with the post-X
>>>
>>> Thinking about some kind of pause... This way at least we won't break
>>> the lifecycle and leave it clean.
>>>
>>
>> That’s easy. Have a Maven-pause-plugin that just waits for you to press
>> enter. Bind it to integration-test in a profile and presto!
>>
>> But that removes the need for the current explicit phases of pre- and
>> post-
>>
>> TBH I think we need to lay down the plan that we want to go towards. It
>> will take a while to change existing phases, in part because removing
>> phases is a breaking change. You can have 3rd part plugins that bind
>> executions to multiple phases, expecting those phases to both exist and
>> have specific execution behaviour.
>>
>> Hence why I think we should go all the way technically, but leave the
>> lifecycle mostly as-is (modulo adding any new phases and flagging existing
>> phases as deprecated).
>>
>> Half measures will only prolong to pain for users.
>>
>> If instead we say: “here’s where we were, here’s where we’re going and
>> this is how we get there” people can incorporate that and adapt
>>
>> Messing about with one phase, that’s just hacks. Adding the ability to
>> define phase execution guarantees... that’s where we want to go. Adding the
>> ability to control plugin execution order within phases... that’s where we
>> want to go... is the syntax where we want to go? Probably not, but it’s how
>> we can get there
>>
>>>
>>>
>>> On 15-11-2019 11:07:23, Stephen Connolly <
>>> [hidden email]> wrote:
>>> On Fri 15 Nov 2019 at 09:18, Robert Scholte wrote:
>>>
>>> > On 13-11-2019 21:46:04, Stephen Connolly
>>> > wrote:
>>> > On Wed 13 Nov 2019 at 19:29, Robert Scholte wrote:
>>> >
>>> > > The name of the branch contains MNG-5668, but it contains much more.
>>> > > I'd likely lead to comments like "great", without being explicit
>>> saying
>>> > > which part(s).
>>> > > I am aware there's all proposals touch the same code, but can be
>>> released
>>> > > isolated from each other.
>>> > > e.g. if the enums-value are changed to "pre-" and "post-" it should
>>> work
>>> > > for the existing phases, which means we could already use it quite
>>> soon
>>> > > (still need to test it myself, though)
>>> > > I also want to provide a counter proposal, but that takes time and
>>> for me
>>> > > there are other issues more important.
>>> >
>>> >
>>> > How would you handle the use case that we’ve already had reported:
>>> >
>>> > As a user I want to test my integration tests in my IDE by running `mvn
>>> > integration-test` so that the test environment is not torn down and I
>>> can
>>> > debug and rerun the tests until I’m ready
>>> >
>>> > Robert Scholte:
>>> > I'd say if they want to set up there environment for the integration
>>> > tests, they'd be running pre-integration-test.
>>> > Next select in the IDE the test to execute. I don't see an issue here.
>>> > Calling pre-integration-test implies NOT running post-integration-test.
>>>
>>>
>>> I disagree. I think if you run the pre- phase then you should have the
>>> post- also run
>>>
>>> I think we could have a differential failure mode in the pre-phases
>>> though.
>>> Iow a pre- phase failure returns a different exit code than the actual
>>> phase itself
>>>
>>> >
>>> >
>>> > Every time I explain people about how Maven works with phases, they are
>>> > amazed it doesn't run the post-phase. I doubt we'll see issues if we
>>> switch
>>> > to expected behavior.
>>> >
>>> > Based on the different views, I hope to see more involvement of PMC
>>> > members, because this will be a turning point that probable cannot be
>>> > undone.
>>> >
>>> >
>>> > With the new phases, the existing pom will still work, and some user
>>> opting
>>> > into after:integration-test knows what they are getting
>>> >
>>> >
>>> > >
>>> > > My biggest fear is that this will result in an All-Or-Nothing, and I
>>> like
>>> > > to prevent that. If the try-finally part works as expected we can
>>> extract
>>> > > that part and prepare for one of the next Maven releases.
>>> >
>>> >
>>> > I’d like to understand your fear better. I’ve been playing with the
>>> PoC a
>>> > bit, and TBH it just feels right.
>>> >
>>> > For sure I’d prefer a schema change to encoding in a string, but I’m
>>> also
>>> > inclined towards string encoded dependency GAVs for 5.x so that
>>> wouldn’t be
>>> > the worst if we went that way.
>>> >
>>> > With pom rewriting, I think we could do a 4.1.0 model version that
>>> moved
>>> > the execution point and priority to attributes, by writing as a 4.0.0
>>> with
>>> > the string encoded form... iow rewriting in 4.x allows us to tidy up
>>> the
>>> > schema as long as it has a 1:1 mapping to a 4.0.0 modelVersion that
>>> gets
>>> > deployed.
>>> >
>>> >
>>> > >
>>> > > Robert
>>> > >
>>> > >
>>> > >
>>> > >
>>> > >
>>> > > On 12-11-2019 10:25:42, Stephen Connolly
>>> > > wrote:
>>> > > On Tue 12 Nov 2019 at 07:34, Robert Scholte wrote:
>>> > >
>>> > > > This is not just MNG-5668, but also contains several non-existing
>>> > issues,
>>> > > > that should be mentioned explicitly as they will have huge impact:
>>> > > >
>>> > > > - support before:/after: prefix for phase-binding
>>> > > >
>>> > > > - introduce priority
>>> > > > - reduce phases (this one hasn't been implemented, but seems to be
>>> the
>>> > > > reason behind before:/after:)
>>> > >
>>> > >
>>> > > All detailed in the proposal on the wiki:
>>> > > https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases
>>> > >
>>> > > Reducing phases would be a big change and not before 4.x at least
>>> (maybe
>>> > > 5.x more realistically... at least we’d need to deprecate the phases
>>> for
>>> > a
>>> > > good while before removing any)
>>> > >
>>> > >
>>> > > >
>>> > > > I would like see separate branches for all of them, as they all
>>> have
>>> > > their
>>> > > > own discussion.
>>> > >
>>> > >
>>> > > The whole point of a PoC is the get feedback. I don’t see utility in
>>> > > separate branches as they are all touching the same code.
>>> > >
>>> > > Once we get feedback we can decide where we want to go from there.
>>> > >
>>> > >
>>> > > >
>>> > > > Robert
>>> > > > On 11-11-2019 20:31:44, Stephen Connolly
>>> > > > wrote:
>>> > > > https://github.com/apache/maven/tree/mng-5668-poc is my POC
>>> > > implementation
>>> > > > for anyone interested in trying it out.
>>> > > >
>>> > > > Here's a pom that builds with the PoC
>>> > > >
>>> > > >
>>> > > > 4.0.0
>>> > > > localdomain
>>> > > > foo
>>> > > > 1.0-SNAPSHOT
>>> > > >
>>> > > >
>>> > > >
>>> > > > maven-antrun-plugin
>>> > > >
>>> > > >
>>> > > > 1
>>> > > > before:integration-test
>>> > > >
>>> > > > run
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > > 2
>>> > > > before:integration-test[1000]
>>> > > >
>>> > > > run
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > > 3
>>> > > > after:integration-test
>>> > > >
>>> > > > run
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > > 4
>>> > > > integration-test
>>> > > >
>>> > > > run
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > >
>>> > > > On Sun, 27 Oct 2019 at 10:55, Robert Scholte wrote:
>>> > > >
>>> > > > > TLDR: We can do better than, but who is in control?
>>> lifecycle-owner,
>>> > > > > plugin-owner or pom-owner?
>>> > > > >
>>> > > > > I think we all recognize the issues we're trying to solve, but
>>> to me
>>> > > this
>>> > > > > proposal is not the right solution.
>>> > > > >
>>> > > > > In general there are 2 issues:
>>> > > > > 1. provide a mechanism that makes sure some executions are called
>>> > even
>>> > > > its
>>> > > > > matching main phase fails.
>>> > > > > 2. provide a mechanism then ensures the order of executions.
>>> > > > >
>>> > > > > The problem of issue 1 is described in MNG-5668, but not the
>>> final
>>> > > > > solution.
>>> > > > > MNG-5668 proposes to give this power to the *lifecycle-owner*,
>>> > whereas
>>> > > > > stage 2 proposes to give the power to the *pom-owner*.
>>> > > > > Both agree on the same thing: by default these post-phases
>>> should be
>>> > > > > triggered even after failure of the matching main phase. This is
>>> > > actually
>>> > > > > already expected behavior, so I don't expect real issues when
>>> > > > implementing
>>> > > > > this adjusted behavior.
>>> > > > > To me after:integration-test is just an alias for
>>> > > post-integration-test,
>>> > > > > both should work the same way.
>>> > > > >
>>> > > > > Issue 2 is a more common problem: controlling the order of
>>> > executions.
>>> > > > > In some cases it is pretty hard or even impossible to get the
>>> > preferred
>>> > > > > order. The latter happens when 2 goals of the same plugin must be
>>> > > > executed
>>> > > > > and a goal of another plugin are competing within the same phase.
>>> > > > >
>>> > > > > So let's first take a look at a phase: is there a clear
>>> definition?
>>> > > > > "A phase is a step in what Maven calls a 'build lifecycle'. The
>>> build
>>> > > > > lifecycle is an ordered sequence of phases involved in building a
>>> > > > project".
>>> > > > > "Lifecycle phases are intentionally vague, defined solely as
>>> > > > > validation, testing, or deployment, and they may mean different
>>> > things
>>> > > to
>>> > > > > different projects."
>>> > > > > Phases are intended to be called from the commandline, and
>>> within the
>>> > > pom
>>> > > > > you define you can control what should happen before or during
>>> that
>>> > > > phase.
>>> > > > >
>>> > > > > To me changing the content of the -element is a codesmell as it
>>> > > > > becomes more than just a phase, and we start programming. Why do
>>> we
>>> > > need
>>> > > > it?
>>> > > > > In the end it is all about ensuring the order of plugin
>>> executions.
>>> > > > > Stage3+4 proposes to give the power to the *pom-owner*,
>>> > > > > whereas MPLUGIN-350[2] proposes to give this power to the
>>> > > *plugin-owner*.
>>> > > > > IIUR Gradle does not have this issue, because their plugins are
>>> aware
>>> > > of
>>> > > > > input and output. They ensure that if the output plugin X is the
>>> > input
>>> > > of
>>> > > > > plugin Y, than X is executed before Y.
>>> > > > > And we should do the same. And this comes with benefits: we can
>>> > decide
>>> > > if
>>> > > > > executions within a project can be executed in parallel. And the
>>> pom
>>> > > > stays
>>> > > > > as clean as it is right now.
>>> > > > >
>>> > > > > In cases when there's a better ownership than the pom-owner, I
>>> would
>>> > > > > prefer to choose that solution. I already notice how people
>>> (don't)
>>> > > build
>>> > > > > up their knowledge regarding poms. The lifecycle-owner and
>>> > plugin-owner
>>> > > > > know much better what they're doing.
>>> > > > >
>>> > > > > thanks,
>>> > > > > Robert
>>> > > > >
>>> > > > > Some food for thoughts: consider a developer that wants to run up
>>> > until
>>> > > > > pre-integration-test, because he wants to bring his system in a
>>> > certain
>>> > > > > state so he can work with IDE to do some work.Can we say that If
>>> And
>>> > > Only
>>> > > > > If somebody called the pre-PHASE, there's no reason to end with
>>> the
>>> > > > > post-PHASE?
>>> > > > >
>>> > > > > [1] https://issues.apache.org/jira/browse/MNG-5668
>>> > > > > [2] https://issues.apache.org/jira/browse/MPLUGIN-350
>>> > > > > On 26-10-2019 14:20:50, Stephen Connolly
>>> > > > > wrote:
>>> > > > > On Sat 26 Oct 2019 at 10:50, Robert Scholte wrote:
>>> > > > >
>>> > > > > > To avoid confusion, let's call it stages.
>>> > > > > >
>>> > > > > > Stage 1: Always call post-bound executions (MNG-5665[1])
>>> > > > > > Stage 2: before and after
>>> > > > > > Stage 3: priorities (MNG-3522[2])
>>> > > > > > Stage 4: transitional lifecycle
>>> > > > >
>>> > > > >
>>> > > > > I have a prototype of stages 1-3 nearly (80%) done... just have
>>> to
>>> > > polish
>>> > > > > up and validate the bound executions with some tests
>>> > > > >
>>> > > > >
>>> > > > > >
>>> > > > > > For both all you need to start evaluating the value of phase.
>>> > > > > > For now we can assume that after:clean is just another label
>>> for
>>> > > > > > post-clean and will have exactly the same effect.
>>> > > > > > MNG-5665 contains a proposal to change the xml, but we
>>> shouldn't do
>>> > > > that
>>> > > > > > (yet). Let's start with a hardcoded list of postphases (or in
>>> case
>>> > a
>>> > > > goal
>>> > > > > > fails, see if a post-x phase exists). Stage 1 is to make it
>>> work,
>>> > > > stage 2
>>> > > > > > to make it configurable.
>>> > > > > > IIRC you cannot ask from inside a Mojo if is was called
>>> explicitly
>>> > or
>>> > > > > > because it was bound to a phase, nor can you ask for the value
>>> of
>>> > > this
>>> > > > > > phase. I kind of like this, plugins shouldn't care about this.
>>> > > > > > However, inside Maven it will become important at which phase
>>> it is
>>> > > to
>>> > > > > > know if there are more executions to call OR create blocks of
>>> > > > executions.
>>> > > > > > Now it is just a list of executions: loop and fail fast.
>>> > > > > >
>>> > > > > > thanks,
>>> > > > > > Robert
>>> > > > > >
>>> > > > > > [1] https://issues.apache.org/jira/browse/MNG-5665
>>> > > > > > [2] https://issues.apache.org/jira/browse/MNG-3522
>>> > > > > > On 25-10-2019 21:33:14, Stephen Connolly
>>> > > > > > wrote:
>>> > > > > > Robert,
>>> > > > > >
>>> > > > > > I would be fine splitting out into, pardon the pun, phases:
>>> > > > > >
>>> > > > > > Phase 1: before and after
>>> > > > > > Phase 2: priorities
>>> > > > > > Phase 3: transitional lifecycle
>>> > > > > >
>>> > > > > > Might have a phase 1.5 of before:* and after:* to catch the
>>> start
>>> > of
>>> > > a
>>> > > > > > lifecycle and the end of a lifecycle...
>>> > > > > >
>>> > > > > > On Fri 25 Oct 2019 at 20:30, Stephen Connolly
>>> > > > > > [hidden email] [mailto:
>>> > > > [hidden email]
>>> > > > > ]>
>>> > > > > > wrote:
>>> > > > > >
>>> > > > > > Robert, Michael, Tibor, let’s continue here (though I asked
>>> Infra
>>> > and
>>> > > > > it’s
>>> > > > > > fine that anyone in the community can join our Slack)
>>> > > > > >
>>> > > > > > On Fri 25 Oct 2019 at 20:01, Stephen Connolly
>>> > > > > > [hidden email] [mailto:
>>> > > > [hidden email]
>>> > > > > ]>
>>> > > > > > wrote:
>>> > > > > >
>>> > > > > >
>>> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases [
>>> > > > > >
>>> https://cwiki.apache.org/confluence/display/MAVEN/Dynamic+phases]
>>> > > > > >
>>> > > > > > Thoughts?
>>> > > > > > --
>>> > > > > >
>>> > > > > > Sent from my phone
>>> > > > > > --
>>> > > > > >
>>> > > > > > Sent from my phone
>>> > > > > > --
>>> > > > > >
>>> > > > > > Sent from my phone
>>> > > > >
>>> > > > > --
>>> > > > > Sent from my phone
>>> > > > >
>>> > > >
>>> > > --
>>> > > Sent from my phone
>>> > >
>>> > --
>>> > Sent from my phone
>>> >
>>> --
>>> Sent from my phone
>>>
>> --
>> Sent from my phone
>>
>