You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@maven.apache.org by Hervé Boutemy <he...@free.fr> on 2024/03/24 15:48:17 UTC
release of maven-source-plugin
I'd like to release maven-source-plugin 3.3.1 to drop the umask sensitivity
during Reproducible Builds :)1
but there are a few issues open
https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
they are related to https://issues.apache.org/jira/browse/MSOURCES-121: I
don't get what conditions made the build fail previously, but MSOURCES-121
make it fail now always.
Should we just change the ERROR into WARNING?
Should we do something smarter, for example not re-add if it's the same file?
Or warn only if the second file addition is different from the first? Then
replace instead of add?
Please help clarifying and fixing this issue that seems ignored for a long time
Regards,
Hervé
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: release of maven-source-plugin
Posted by Herve Boutemy <hb...@apache.org>.
https://github.com/apache/maven-source-plugin/pull/24 does not seem to get any positive traction: be it
MSOURCE-140 and MSOURCE-141 are assigned to 3.3.1 release but nobody wants to work with me on it to solve something: it seems they are not a real problem
I won't hold the 3.3.1 release forever, I'll do it without them fixed to benefit from other fixes: if someone wants to work on it for future releases, I'll let them do
release early, release often, one step at a time
On 2024/03/25 07:10:22 Hervé Boutemy wrote:
> PR ready for review:
> https://github.com/apache/maven-source-plugin/pull/24
>
> I think it does the right job at differentiating 2 cases that the initial
> MSOURCES-121 merged too aggressively:
> 1. do not fail when executed twice with the same output file
> 2. but fail when re-attach happens to another output file = another classifier
> has to be configured
>
> Regards,
>
> Hervé
>
> Le lundi 25 mars 2024, 07:31:53 CET Hervé Boutemy a écrit :
> > IIUC, this is an explanation of the mystery failure that started the
> > MSOURCE-121 update
> >
> > but after MSOURCE-121, the plugin itself stops with a failure when detecting
> > re-addition: that's why Gary can't upgrade, the failure is now happening at
> > plugin level, and always
> > "mvn install deploy" (with source enabled) is a simple test to see the
> > failure introduced by MSOURCE-121
> >
> > I'm starting to understand the many aspects of this issue: I think I know
> > how to update the goal to be more tolerant
> >
> > Thanks for the help
> >
> > Hervé
> >
> > Le dimanche 24 mars 2024, 20:39:55 CET Romain Manni-Bucau a écrit :
> > > Hi,
> > >
> > > I think it is fixed for v4 serie.
> > > Main issue comes from the default v3 pom (
> > > https://github.com/apache/maven/blob/maven-3.9.2/maven-model-builder/src/m
> > > ai n/resources/org/apache/maven/model/pom-4.0.0.xml#L113) which is no more
> > > a thing in v4 so the merge of the plugin executions does not happen the
> > > same and the issue disappeared.
> > > The easiest is likely to either let the default be merged and not use
> > > "jar"
> > > (implicit jar-no-fork instead) or just not use the default id
> > > (attach-sources) and skip the default one.
> > >
> > > So there is no source plugin issue (as the plugin will never fix it by
> > > "bug
> > > design") so no reason to hold a release IMHO.
> > >
> > > Guess it silently overrided the already built artifact for years - until
> > > we
> > > just fail cause it is a broken design - after this change of goal
> > > https://issues.apache.org/jira/browse/MNG-5940.
> > >
> > > So long story short: we are on track and release can be done I think.
> > >
> > > Romain Manni-Bucau
> > > @rmannibucau <https://twitter.com/rmannibucau> | Blog
> > > <https://rmannibucau.metawerx.net/> | Old Blog
> > > <http://rmannibucau.wordpress.com> | Github
> > > <https://github.com/rmannibucau>>
> > > | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book
> > >
> > > <https://www.packtpub.com/application-development/java-ee-8-high-performan
> > > ce
> > >
> > >
> > >
> > > Le dim. 24 mars 2024 à 18:44, Gary Gregory <ga...@gmail.com> a
> > >
> > > écrit :
> > > > Hi All,
> > > >
> > > > As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
> > > > play, Commons will stay on a pre-3.3.0 version.
> > > >
> > > > I re-read the comments, I still don't know what we can do in Commons
> > > > to address this as this feels like some deep Maven Magic.
> > > >
> > > > Gary
> > > >
> > > > On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr>
> > > >
> > > > wrote:
> > > > > I'd like to release maven-source-plugin 3.3.1 to drop the umask
> > > >
> > > > sensitivity
> > > >
> > > > > during Reproducible Builds :)1
> > > > >
> > > > > but there are a few issues open
> > > > >
> > > > > https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
> > > > >
> > > > > they are related to https://issues.apache.org/jira/browse/
> MSOURCES-121:
> > > > I
> > > >
> > > > > don't get what conditions made the build fail previously, but
> > > >
> > > > MSOURCES-121
> > > >
> > > > > make it fail now always.
> > > > > Should we just change the ERROR into WARNING?
> > > > > Should we do something smarter, for example not re-add if it's the
> > > > > same
> > > >
> > > > file?
> > > >
> > > > > Or warn only if the second file addition is different from the first?
> > > >
> > > > Then
> > > >
> > > > > replace instead of add?
> > > > >
> > > > > Please help clarifying and fixing this issue that seems ignored for a
> > > >
> > > > long time
> > > >
> > > > > Regards,
> > > > >
> > > > > Hervé
> > > > >
> > > > >
> > > > >
> > > > > ---------------------------------------------------------------------
> > > > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > > > For additional commands, e-mail: dev-help@maven.apache.org
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > > For additional commands, e-mail: dev-help@maven.apache.org
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > For additional commands, e-mail: dev-help@maven.apache.org
>
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: release of maven-source-plugin
Posted by Hervé Boutemy <he...@free.fr>.
PR ready for review:
https://github.com/apache/maven-source-plugin/pull/24
I think it does the right job at differentiating 2 cases that the initial
MSOURCES-121 merged too aggressively:
1. do not fail when executed twice with the same output file
2. but fail when re-attach happens to another output file = another classifier
has to be configured
Regards,
Hervé
Le lundi 25 mars 2024, 07:31:53 CET Hervé Boutemy a écrit :
> IIUC, this is an explanation of the mystery failure that started the
> MSOURCE-121 update
>
> but after MSOURCE-121, the plugin itself stops with a failure when detecting
> re-addition: that's why Gary can't upgrade, the failure is now happening at
> plugin level, and always
> "mvn install deploy" (with source enabled) is a simple test to see the
> failure introduced by MSOURCE-121
>
> I'm starting to understand the many aspects of this issue: I think I know
> how to update the goal to be more tolerant
>
> Thanks for the help
>
> Hervé
>
> Le dimanche 24 mars 2024, 20:39:55 CET Romain Manni-Bucau a écrit :
> > Hi,
> >
> > I think it is fixed for v4 serie.
> > Main issue comes from the default v3 pom (
> > https://github.com/apache/maven/blob/maven-3.9.2/maven-model-builder/src/m
> > ai n/resources/org/apache/maven/model/pom-4.0.0.xml#L113) which is no more
> > a thing in v4 so the merge of the plugin executions does not happen the
> > same and the issue disappeared.
> > The easiest is likely to either let the default be merged and not use
> > "jar"
> > (implicit jar-no-fork instead) or just not use the default id
> > (attach-sources) and skip the default one.
> >
> > So there is no source plugin issue (as the plugin will never fix it by
> > "bug
> > design") so no reason to hold a release IMHO.
> >
> > Guess it silently overrided the already built artifact for years - until
> > we
> > just fail cause it is a broken design - after this change of goal
> > https://issues.apache.org/jira/browse/MNG-5940.
> >
> > So long story short: we are on track and release can be done I think.
> >
> > Romain Manni-Bucau
> > @rmannibucau <https://twitter.com/rmannibucau> | Blog
> > <https://rmannibucau.metawerx.net/> | Old Blog
> > <http://rmannibucau.wordpress.com> | Github
> > <https://github.com/rmannibucau>>
> > | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book
> >
> > <https://www.packtpub.com/application-development/java-ee-8-high-performan
> > ce
> >
> >
> >
> > Le dim. 24 mars 2024 à 18:44, Gary Gregory <ga...@gmail.com> a
> >
> > écrit :
> > > Hi All,
> > >
> > > As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
> > > play, Commons will stay on a pre-3.3.0 version.
> > >
> > > I re-read the comments, I still don't know what we can do in Commons
> > > to address this as this feels like some deep Maven Magic.
> > >
> > > Gary
> > >
> > > On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr>
> > >
> > > wrote:
> > > > I'd like to release maven-source-plugin 3.3.1 to drop the umask
> > >
> > > sensitivity
> > >
> > > > during Reproducible Builds :)1
> > > >
> > > > but there are a few issues open
> > > >
> > > > https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
> > > >
> > > > they are related to https://issues.apache.org/jira/browse/
MSOURCES-121:
> > > I
> > >
> > > > don't get what conditions made the build fail previously, but
> > >
> > > MSOURCES-121
> > >
> > > > make it fail now always.
> > > > Should we just change the ERROR into WARNING?
> > > > Should we do something smarter, for example not re-add if it's the
> > > > same
> > >
> > > file?
> > >
> > > > Or warn only if the second file addition is different from the first?
> > >
> > > Then
> > >
> > > > replace instead of add?
> > > >
> > > > Please help clarifying and fixing this issue that seems ignored for a
> > >
> > > long time
> > >
> > > > Regards,
> > > >
> > > > Hervé
> > > >
> > > >
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > > For additional commands, e-mail: dev-help@maven.apache.org
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > For additional commands, e-mail: dev-help@maven.apache.org
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: release of maven-source-plugin
Posted by Hervé Boutemy <he...@free.fr>.
IIUC, this is an explanation of the mystery failure that started the
MSOURCE-121 update
but after MSOURCE-121, the plugin itself stops with a failure when detecting
re-addition: that's why Gary can't upgrade, the failure is now happening at
plugin level, and always
"mvn install deploy" (with source enabled) is a simple test to see the failure
introduced by MSOURCE-121
I'm starting to understand the many aspects of this issue: I think I know how
to update the goal to be more tolerant
Thanks for the help
Hervé
Le dimanche 24 mars 2024, 20:39:55 CET Romain Manni-Bucau a écrit :
> Hi,
>
> I think it is fixed for v4 serie.
> Main issue comes from the default v3 pom (
> https://github.com/apache/maven/blob/maven-3.9.2/maven-model-builder/src/mai
> n/resources/org/apache/maven/model/pom-4.0.0.xml#L113) which is no more a
> thing in v4 so the merge of the plugin executions does not happen the same
> and the issue disappeared.
> The easiest is likely to either let the default be merged and not use "jar"
> (implicit jar-no-fork instead) or just not use the default id
> (attach-sources) and skip the default one.
>
> So there is no source plugin issue (as the plugin will never fix it by "bug
> design") so no reason to hold a release IMHO.
>
> Guess it silently overrided the already built artifact for years - until we
> just fail cause it is a broken design - after this change of goal
> https://issues.apache.org/jira/browse/MNG-5940.
>
> So long story short: we are on track and release can be done I think.
>
> Romain Manni-Bucau
> @rmannibucau <https://twitter.com/rmannibucau> | Blog
> <https://rmannibucau.metawerx.net/> | Old Blog
> <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau>
> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book
> <https://www.packtpub.com/application-development/java-ee-8-high-performance
> >
>
>
> Le dim. 24 mars 2024 à 18:44, Gary Gregory <ga...@gmail.com> a
>
> écrit :
> > Hi All,
> >
> > As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
> > play, Commons will stay on a pre-3.3.0 version.
> >
> > I re-read the comments, I still don't know what we can do in Commons
> > to address this as this feels like some deep Maven Magic.
> >
> > Gary
> >
> > On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr>
> >
> > wrote:
> > > I'd like to release maven-source-plugin 3.3.1 to drop the umask
> >
> > sensitivity
> >
> > > during Reproducible Builds :)1
> > >
> > > but there are a few issues open
> > >
> > > https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
> > >
> > > they are related to https://issues.apache.org/jira/browse/MSOURCES-121:
> > I
> >
> > > don't get what conditions made the build fail previously, but
> >
> > MSOURCES-121
> >
> > > make it fail now always.
> > > Should we just change the ERROR into WARNING?
> > > Should we do something smarter, for example not re-add if it's the same
> >
> > file?
> >
> > > Or warn only if the second file addition is different from the first?
> >
> > Then
> >
> > > replace instead of add?
> > >
> > > Please help clarifying and fixing this issue that seems ignored for a
> >
> > long time
> >
> > > Regards,
> > >
> > > Hervé
> > >
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > For additional commands, e-mail: dev-help@maven.apache.org
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > For additional commands, e-mail: dev-help@maven.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: release of maven-source-plugin
Posted by Gary Gregory <ga...@gmail.com>.
I'm glad to hear there is nothing fundamentally wrong. Thank you for the
detailed explanation.
Gary
On Sun, Mar 24, 2024, 3:40 PM Romain Manni-Bucau <rm...@gmail.com>
wrote:
> Hi,
>
> I think it is fixed for v4 serie.
> Main issue comes from the default v3 pom (
>
> https://github.com/apache/maven/blob/maven-3.9.2/maven-model-builder/src/main/resources/org/apache/maven/model/pom-4.0.0.xml#L113
> )
> which is no more a thing in v4 so the merge of the plugin executions does
> not happen the same and the issue disappeared.
> The easiest is likely to either let the default be merged and not use "jar"
> (implicit jar-no-fork instead) or just not use the default id
> (attach-sources) and skip the default one.
>
> So there is no source plugin issue (as the plugin will never fix it by "bug
> design") so no reason to hold a release IMHO.
>
> Guess it silently overrided the already built artifact for years - until we
> just fail cause it is a broken design - after this change of goal
> https://issues.apache.org/jira/browse/MNG-5940.
>
> So long story short: we are on track and release can be done I think.
>
> Romain Manni-Bucau
> @rmannibucau <https://twitter.com/rmannibucau> | Blog
> <https://rmannibucau.metawerx.net/> | Old Blog
> <http://rmannibucau.wordpress.com> | Github <
> https://github.com/rmannibucau> |
> LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book
> <
> https://www.packtpub.com/application-development/java-ee-8-high-performance
> >
>
>
> Le dim. 24 mars 2024 à 18:44, Gary Gregory <ga...@gmail.com> a
> écrit :
>
> > Hi All,
> >
> > As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
> > play, Commons will stay on a pre-3.3.0 version.
> >
> > I re-read the comments, I still don't know what we can do in Commons
> > to address this as this feels like some deep Maven Magic.
> >
> > Gary
> >
> > On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr>
> > wrote:
> > >
> > > I'd like to release maven-source-plugin 3.3.1 to drop the umask
> > sensitivity
> > > during Reproducible Builds :)1
> > >
> > > but there are a few issues open
> > > https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
> > >
> > > they are related to https://issues.apache.org/jira/browse/MSOURCES-121
> :
> > I
> > > don't get what conditions made the build fail previously, but
> > MSOURCES-121
> > > make it fail now always.
> > > Should we just change the ERROR into WARNING?
> > > Should we do something smarter, for example not re-add if it's the same
> > file?
> > > Or warn only if the second file addition is different from the first?
> > Then
> > > replace instead of add?
> > >
> > > Please help clarifying and fixing this issue that seems ignored for a
> > long time
> > >
> > > Regards,
> > >
> > > Hervé
> > >
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > > For additional commands, e-mail: dev-help@maven.apache.org
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > For additional commands, e-mail: dev-help@maven.apache.org
> >
> >
>
Re: release of maven-source-plugin
Posted by Romain Manni-Bucau <rm...@gmail.com>.
Hi,
I think it is fixed for v4 serie.
Main issue comes from the default v3 pom (
https://github.com/apache/maven/blob/maven-3.9.2/maven-model-builder/src/main/resources/org/apache/maven/model/pom-4.0.0.xml#L113)
which is no more a thing in v4 so the merge of the plugin executions does
not happen the same and the issue disappeared.
The easiest is likely to either let the default be merged and not use "jar"
(implicit jar-no-fork instead) or just not use the default id
(attach-sources) and skip the default one.
So there is no source plugin issue (as the plugin will never fix it by "bug
design") so no reason to hold a release IMHO.
Guess it silently overrided the already built artifact for years - until we
just fail cause it is a broken design - after this change of goal
https://issues.apache.org/jira/browse/MNG-5940.
So long story short: we are on track and release can be done I think.
Romain Manni-Bucau
@rmannibucau <https://twitter.com/rmannibucau> | Blog
<https://rmannibucau.metawerx.net/> | Old Blog
<http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> |
LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book
<https://www.packtpub.com/application-development/java-ee-8-high-performance>
Le dim. 24 mars 2024 à 18:44, Gary Gregory <ga...@gmail.com> a
écrit :
> Hi All,
>
> As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
> play, Commons will stay on a pre-3.3.0 version.
>
> I re-read the comments, I still don't know what we can do in Commons
> to address this as this feels like some deep Maven Magic.
>
> Gary
>
> On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr>
> wrote:
> >
> > I'd like to release maven-source-plugin 3.3.1 to drop the umask
> sensitivity
> > during Reproducible Builds :)1
> >
> > but there are a few issues open
> > https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
> >
> > they are related to https://issues.apache.org/jira/browse/MSOURCES-121:
> I
> > don't get what conditions made the build fail previously, but
> MSOURCES-121
> > make it fail now always.
> > Should we just change the ERROR into WARNING?
> > Should we do something smarter, for example not re-add if it's the same
> file?
> > Or warn only if the second file addition is different from the first?
> Then
> > replace instead of add?
> >
> > Please help clarifying and fixing this issue that seems ignored for a
> long time
> >
> > Regards,
> >
> > Hervé
> >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > For additional commands, e-mail: dev-help@maven.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
Re: release of maven-source-plugin
Posted by Gary Gregory <ga...@gmail.com>.
Hi All,
As long as https://issues.apache.org/jira/browse/MSOURCES-143 is in
play, Commons will stay on a pre-3.3.0 version.
I re-read the comments, I still don't know what we can do in Commons
to address this as this feels like some deep Maven Magic.
Gary
On Sun, Mar 24, 2024 at 11:48 AM Hervé Boutemy <he...@free.fr> wrote:
>
> I'd like to release maven-source-plugin 3.3.1 to drop the umask sensitivity
> during Reproducible Builds :)1
>
> but there are a few issues open
> https://issues.apache.org/jira/projects/MSOURCES/versions/12353471
>
> they are related to https://issues.apache.org/jira/browse/MSOURCES-121: I
> don't get what conditions made the build fail previously, but MSOURCES-121
> make it fail now always.
> Should we just change the ERROR into WARNING?
> Should we do something smarter, for example not re-add if it's the same file?
> Or warn only if the second file addition is different from the first? Then
> replace instead of add?
>
> Please help clarifying and fixing this issue that seems ignored for a long time
>
> Regards,
>
> Hervé
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org