[jira] (MRELEASE-431) Configuration of policy for calculating next (release) version

2013-09-17 Thread Robert Munteanu (JIRA)

 [ 
https://jira.codehaus.org/browse/MRELEASE-431?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Robert Munteanu updated MRELEASE-431:
-

Attachment: MRELEASE-431-v3.patch

> Configuration of policy for calculating next (release) version
> --
>
> Key: MRELEASE-431
> URL: https://jira.codehaus.org/browse/MRELEASE-431
> Project: Maven Release Plugin
>  Issue Type: New Feature
>  Components: prepare
>Affects Versions: 2.0-beta-8
>Reporter: Carsten Ziegeler
>Assignee: Robert Scholte
> Fix For: 2.5
>
> Attachments: 
> 0001-MRELEASE-431-Configuration-of-policy-for-calculating.patch, 
> MRELEASE-431.patch, MRELEASE-431-v2.patch, MRELEASE-431-v3.patch
>
>
> Currently, when preparing the release, the version to release is always the 
> next version which usually is the current version without the snapshot 
> extension.
> There are quiet a lot projects (Apache Felix, Sling and others) following an 
> even release numbering policy. So while the current development version is 
> odd (like 1.2.3-SNAPSHOT), the next released version will be 1.2.4.
> It would be nice if this could be made configuration through some 
> configuration property like
> next-even (with possible values being: next 
> (default, as-is), next-even, next-odd
> I briefly scanned through the code and it seems that adding support for this 
> requires changes in both, the release-manager and the release-plugin.
> If this feature gets accepted and if someone could give me some minor hints 
> how/where to add this I could come up with a patch.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] (MRELEASE-431) Configuration of policy for calculating next (release) version

2013-09-17 Thread Robert Munteanu (JIRA)

[ 
https://jira.codehaus.org/browse/MRELEASE-431?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=332905#comment-332905
 ] 

Robert Munteanu commented on MRELEASE-431:
--

[~rfscholte] - I've attached [^MRELEASE-431-v3.patch], with the 
VersionPolicyRequest addition ( [diff on 
github|https://github.com/rombert/release/compare/6de575dd76224c690bfd88375eac88f49ba6d319...HEAD]
 ).

> Configuration of policy for calculating next (release) version
> --
>
> Key: MRELEASE-431
> URL: https://jira.codehaus.org/browse/MRELEASE-431
> Project: Maven Release Plugin
>  Issue Type: New Feature
>  Components: prepare
>Affects Versions: 2.0-beta-8
>Reporter: Carsten Ziegeler
>Assignee: Robert Scholte
> Fix For: 2.5
>
> Attachments: 
> 0001-MRELEASE-431-Configuration-of-policy-for-calculating.patch, 
> MRELEASE-431.patch, MRELEASE-431-v2.patch, MRELEASE-431-v3.patch
>
>
> Currently, when preparing the release, the version to release is always the 
> next version which usually is the current version without the snapshot 
> extension.
> There are quiet a lot projects (Apache Felix, Sling and others) following an 
> even release numbering policy. So while the current development version is 
> odd (like 1.2.3-SNAPSHOT), the next released version will be 1.2.4.
> It would be nice if this could be made configuration through some 
> configuration property like
> next-even (with possible values being: next 
> (default, as-is), next-even, next-odd
> I briefly scanned through the code and it seems that adding support for this 
> requires changes in both, the release-manager and the release-plugin.
> If this feature gets accepted and if someone could give me some minor hints 
> how/where to add this I could come up with a patch.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] (MRELEASE-594) release:prepare should stop when there is snapshots in dependencies management

2013-09-17 Thread Konrad Foegen (JIRA)

 [ 
https://jira.codehaus.org/browse/MRELEASE-594?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Konrad Foegen updated MRELEASE-594:
---

Attachment: allowUnusedSnapshotsInDependencyManagement.patch

Hi,

please have a look at the patch I've attached. By default the behaviour of the 
plugin remains the same but it can be "enabled" by adding the following flag 
-DallowUnusedSnapshotsInDependencyManagement=false (the default is true).

If enabled, it treats snapshots in the dependency management section the same 
way as snapshots in the dependency section, i.e. snapshots can be resolved or 
the release fails. 

> release:prepare should stop when there is snapshots in dependencies management
> --
>
> Key: MRELEASE-594
> URL: https://jira.codehaus.org/browse/MRELEASE-594
> Project: Maven Release Plugin
>  Issue Type: Bug
>  Components: prepare
>Reporter: Thomas Bruyelle
>Assignee: Olivier Lamy
> Fix For: Backlog
>
> Attachments: allowUnusedSnapshotsInDependencyManagement.patch
>
>


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] (MRELEASE-594) release:prepare should stop when there is snapshots in dependencies management

2013-09-17 Thread Hans Aikema (JIRA)

[ 
https://jira.codehaus.org/browse/MRELEASE-594?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=332916#comment-332916
 ] 

Hans Aikema commented on MRELEASE-594:
--

I agree with previous commenters that DISallowing snapshots in dependencies in 
the release of any project should be the default (breaking backward stability 
in favor of forward build stability as no-one can predict what will change in 
the snapshot version after the release of the POM, making projects that 
safeguard their maintenance on 'latest released' version of a parent still run 
the risk of breaking builds/releases). A released parent POM should by default 
give a stable development environment to build and release on. A snapshot 
parent POM may introduce all kinds of dependency-caused breaking software.
Snapshot-versions of dependencies introduced by a released parent-POM breaking 
your release cycle is completely counter-intuitive to the meaning of stable 
artefacts in Maven.

> release:prepare should stop when there is snapshots in dependencies management
> --
>
> Key: MRELEASE-594
> URL: https://jira.codehaus.org/browse/MRELEASE-594
> Project: Maven Release Plugin
>  Issue Type: Bug
>  Components: prepare
>Reporter: Thomas Bruyelle
>Assignee: Olivier Lamy
> Fix For: Backlog
>
> Attachments: allowUnusedSnapshotsInDependencyManagement.patch
>
>


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] (MRELEASE-594) release:prepare should stop when there is snapshots in dependencies management

2013-09-17 Thread Hans Aikema (JIRA)

[ 
https://jira.codehaus.org/browse/MRELEASE-594?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=332916#comment-332916
 ] 

Hans Aikema edited comment on MRELEASE-594 at 9/17/13 9:23 AM:
---

I agree with previous commenters that DISallowing snapshots in dependencies in 
the release of any project should be the default (breaking backward 
compatibility in favor of forward build stability as no-one can predict what 
will change in the snapshot version after the release of the POM, making 
projects that safeguard their maintenance on 'latest released' version of a 
parent still run the risk of breaking builds/releases). A released parent POM 
should by default give a stable development environment to build and release 
on. A snapshot parent POM may introduce all kinds of dependency-caused breaking 
software.
Snapshot-versions of dependencies introduced by a released parent-POM breaking 
your release cycle is completely counter-intuitive to the meaning of stable 
artefacts in Maven.

  was (Author: aikebah):
I agree with previous commenters that DISallowing snapshots in dependencies 
in the release of any project should be the default (breaking backward 
stability in favor of forward build stability as no-one can predict what will 
change in the snapshot version after the release of the POM, making projects 
that safeguard their maintenance on 'latest released' version of a parent still 
run the risk of breaking builds/releases). A released parent POM should by 
default give a stable development environment to build and release on. A 
snapshot parent POM may introduce all kinds of dependency-caused breaking 
software.
Snapshot-versions of dependencies introduced by a released parent-POM breaking 
your release cycle is completely counter-intuitive to the meaning of stable 
artefacts in Maven.
  
> release:prepare should stop when there is snapshots in dependencies management
> --
>
> Key: MRELEASE-594
> URL: https://jira.codehaus.org/browse/MRELEASE-594
> Project: Maven Release Plugin
>  Issue Type: Bug
>  Components: prepare
>Reporter: Thomas Bruyelle
>Assignee: Olivier Lamy
> Fix For: Backlog
>
> Attachments: allowUnusedSnapshotsInDependencyManagement.patch
>
>


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


Maven parsing pom.xml incorrectly

2013-09-17 Thread Andrew Pennebaker
I'm using Thrift in my Maven project, compiling my .thrift code to .java as
part of the generate-sources step. To do this, I use the maven antrun
plugin in my pom.xml, which executes a command line call to the thrift
executable, and sends it the appropriate argument flags, such as "-out" and
"--gen".

However, when I comment out this plugin with the standard XML comment
syntax , Maven fails to parse the pom file. It doesn't like the
fact that the comment contains --gen, expecting the comment to end right
there.

Can we please improve Maven's comment syntax parsing for pom.xml?


[jira] (MJARSIGNER-13) signing in multi-module project fails on windows

2013-09-17 Thread Chris Feldhacker (JIRA)

[ 
https://jira.codehaus.org/browse/MJARSIGNER-13?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=332937#comment-332937
 ] 

Chris Feldhacker commented on MJARSIGNER-13:


Possibly related tidbit to consider:
We have Windows scripts to install various IDEs and local servers, which 
involves using keytool.exe to import needed certificates into keystores.  When 
we ran the imports back-to-back, it would fail with files in use (sometime it 
would just hang).

What we found was we had to impose a 3-5 second delay between commands before 
they would work.  Don't know if its a Windows issue or perhaps an issue with 
some of the JDK tools, but perhaps something similar is happening here...

Would be nice to at least have the option to NOT rename the file to .orig 
before signing...

> signing in multi-module project fails on windows
> 
>
> Key: MJARSIGNER-13
> URL: https://jira.codehaus.org/browse/MJARSIGNER-13
> Project: Maven Jar Signer Plugin
>  Issue Type: Bug
>Affects Versions: 1.2
> Environment: Windows XP, android sdk 
>Reporter: Anna Gadomska
> Attachments: pom.xml
>
>
> I got multi-module (6 modules) android project with .pom file and "sign" 
> profile defined. When I execute with 'sign' profile:
> - in Maven 3.0.1 - it signs 3 modules and fails. when I resume the execution 
> (mv  -rf) it signs another 3 modules and fails again. The reason for 
> failing is: [INFO] jarsigner: attempt to rename xxx.jar to xxx.jar.orig 
> failed.
> - in Maven 2.2.1 - it doesn't work at all, even for first module; it returns 
> "error code 1"
> However, I tried the same .pom - everything works perfectly in Linux 
> (Ubuntu). 
> I was afraid that it might be the issue with spaces, so I hardcoded the paths 
> - the error I got from maven than was:
> {noformat}
> [ERROR] Failed to execute goal 
> org.apache.maven.plugins:maven-jarsigner-plugin:1.2:sign (signing) on project 
> MyProject: Failed executing 'cmd.exe /X /C 
> "C:\tools\JavaTMSEDevelopementKit\jre\..\bin\jarsigner.exe -verbose -keystore 
> "C:\tmp\debug.keystore" -storepass '*' -keypass '*' xxx.jar 
> '*'debugkey"' - exitcode 1 -> [Help 1]
> {noformat}
> Attaching pom file which I am using.
> Please help?

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] (MSITE-665) Site plugin version 3.2 seems to modify a project's classpath.

2013-09-17 Thread Christian Schulte (JIRA)

[ 
https://jira.codehaus.org/browse/MSITE-665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=332939#comment-332939
 ] 

Christian Schulte commented on MSITE-665:
-

Could someone please perform a release of the site plugin with the attached 
patch applied ? It is a regression and is blocking the upgrade to Maven 3.1.x.


> Site plugin version 3.2 seems to modify a project's classpath.
> --
>
> Key: MSITE-665
> URL: https://jira.codehaus.org/browse/MSITE-665
> Project: Maven Site Plugin
>  Issue Type: Bug
>Affects Versions: 3.2
> Environment: Apache Maven 3.0.4 (r1232337; 2012-01-17 09:44:56+0100)
> Java version: 1.7.0_03, vendor: Oracle Corporation
>Reporter: Christian Schulte
> Attachments: MSITE-665.patch, MSITE-665.zip, msp-3.1.log, msp-3.2.log
>
>


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


bug in parsing file ... comments

2013-09-17 Thread Brian Withnell
We ran into a parsing bug that prevents commenting out an option with a
command line for maven. We tried commenting out (for testing purposes) the
following plugin, which contains the "thrift --gen" command. The "--gen"
caused a parse error in the file (looks like it thinks it needs to have -->
because it is in a comment rather than --> as the only thing that matters.
It would be *great* if it allowed nesting of comments, but I'll take that
it just doesn't barf on valid command structures. Because the following is
a comment, you *should* be able to paste it into any pom file and have
maven process the file properly.