[jira] [Commented] (MPIR-234) SCM-link in site of multimodule projects should not append module name by default (at least for git)

2015-08-13 Thread JIRA

[ 
https://issues.apache.org/jira/browse/MPIR-234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14694875#comment-14694875
 ] 

Hervé Boutemy commented on MPIR-234:


notice that this is what I already did in MPIR-290: display only a part of the 
scm git url in the report
the only weak part is that the algorithm to detect repo part from path part is 
index of '.git'

but this is the update that made me think that we just need a clearly defined 
separator for this info

and everything wil work well: no change in Maven core (keep path behaviour), 
and update in MPIR scm report to better parse the scm url and display parts 
nicely
and of course, update Maven scm to document the new url format (addition of 
this separator to path) and improve the url parser

 SCM-link in site of multimodule projects should not append module name by 
 default (at least for git)
 

 Key: MPIR-234
 URL: https://issues.apache.org/jira/browse/MPIR-234
 Project: Maven Project Info Reports Plugin
  Issue Type: Bug
  Components: scm
Affects Versions: 2.4
Reporter: Mirko Friedenhagen

 I have setup a simple multi module project (see 
 https://github.com/mfriedenhagen/multi-module-sample/tree/multi-site-complex) 
 which uses git on github as {{scm}}. While rendering the site, MPIR will by 
 default add the name of the module to the SCM-URLs in source-repository.html. 
 So instead of https://github.com/mfriedenhagen/multi-module-sample/ I see 
 https://github.com/mfriedenhagen/multi-module-sample/core/, 
 g...@github.com:mfriedenhagen/multi-module-sample.git/core and 
 git://github.com/mfriedenhagen/multi-module-sample.git/core in the report for 
 the core module. All these URLs are invalid. For SVN this could be assumed to 
 be the right behaviour, for git and probably other SCMs this is not true. As 
 a workaround I have to reconfigure the scm section (see 
 https://github.com/mfriedenhagen/multi-module-sample/blob/multi-site-complex/core/pom.xml)
  in the modules like this:
 {code:xml}
 scm
   connection${project.parent.scm.connection}/connection
   
 developerConnection${project.parent.scm.developerConnection}/developerConnection
   url${project.parent.scm.url}/url
 /scm
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org



[GitHub] maven-plugins pull request: Updated docs URL for shade plugin

2015-08-13 Thread edwardmlyte
GitHub user edwardmlyte opened a pull request:

https://github.com/apache/maven-plugins/pull/59

Updated docs URL for shade plugin

Updated due to codehaus repo closing

You can merge this pull request into a Git repository by running:

$ git pull https://github.com/edwardmlyte/maven-plugins trunk

Alternatively you can review and apply these changes as the patch at:

https://github.com/apache/maven-plugins/pull/59.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

This closes #59


commit 58bf43d0dfe31663740f0a480da7da85c7f02e9e
Author: edwardmlyte edwardml...@gmail.com
Date:   2015-08-13T08:52:02Z

Updated docs URL for shade plugin

Updated due to codehaus repo closing




---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org



Read timeout from repository ?

2015-08-13 Thread Kristian Rosenvold
We are discussing (in https://issues.apache.org/jira/browse/MNG-5869)
what appears to be a situation where an intermediate proxy sends an
incorrect content-length and the download seems to wait indefinitely
for the complete payload to await. Is there no timeout configured for
a default get request at all ?

(As Arnaud comments in the issue, it would appear to have been
introduced somewhere after 3.1.1)

Kristian

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org



Re: Read timeout from repository ?

2015-08-13 Thread Olivier Lamy
Hi,
I vaguely remember a discussion [1] regarding a too low RTO of 60s which
was increased to a very high value ([2])

Cheers
--
Olivier
[1] http://markmail.org/message/yqlofeoah6bcxb4i
[2] https://issues.apache.org/jira/browse/WAGON-365


On 13 August 2015 at 21:08, Kristian Rosenvold kristian.rosenv...@gmail.com
 wrote:

 We are discussing (in https://issues.apache.org/jira/browse/MNG-5869)
 what appears to be a situation where an intermediate proxy sends an
 incorrect content-length and the download seems to wait indefinitely
 for the complete payload to await. Is there no timeout configured for
 a default get request at all ?

 (As Arnaud comments in the issue, it would appear to have been
 introduced somewhere after 3.1.1)

 Kristian

 -
 To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
 For additional commands, e-mail: dev-h...@maven.apache.org




-- 
Olivier Lamy
http://twitter.com/olamy | http://linkedin.com/in/olamy


Re: [jira] [Commented] (MPIR-234) SCM-link in site of multimodule projects should not append module name by default (at least for git)

2015-08-13 Thread herve . boutemy
IMHO, what is not clear in your explanations is that scm url is not the url of 
the repo

scm url is scm:provider:provider-specific content

our git scm url is currently weak, since it does not define path in repo: then 
just add it, in a location perfectly disctinct from repo url inside the scm url

then everything will work like svn, even if the feature is not provided by 
the scm itself but by the Maven provider scm url format (which was not 
necessary with svn)

Regards,

Hervé

- Mail original -
De: Lennart Jörelid lennart.jore...@gmail.com
À: Maven Developers List dev@maven.apache.org
Envoyé: Mercredi 12 Août 2015 02:31:33
Objet: Re: [jira] [Commented] (MPIR-234) SCM-link in site of multimodule 
projects should not append module name by default (at least for git)

This is pretty much required whenever one uses a VCS which is not
Subversion.
Everything in Maven having to do with SCMs is very subversion-centric in
that:


   1. By default, maven appends module names (not artifactIDs) to parent
   projects's SCM URLs to form the SCM URL of a child. This does not work for
   any Distributed VCS, as far as i know.
   2. Since the SCM URLs are incorrectly extrapolated, one is forced to
   repeat the calculation in every project, implying quite a lot of Maven
   variables defined in the parents.

It should really be up to the SCM provide to perform all calculation of SCM
URLs. Git URLs are considerably different from SVN or CVS URLs. If the
provider can calulate the effective URL, let it. This would also simplify
the release process; the release plugin could simply query the active SCM
provider for all data it requires.

url${url.prefix}/${path.in.reactor}/url

properties
path.in.reactorthe/actual/vcs/path/not/the/module/names/path.in.reactor
/properties


distributionManagement
site
idsomeId/id
nametheNameOfSomeId/name
url${distro.url.prefix}/${path.in.reactor}/url
/site
/distributionManagement

scm
connection${scm.url.prefix}/${reactor.name}/connection
developerConnection${scm.url.prefix}/${reactor.name}/developerConnection
url${reactor.url.remote.prefix}/${reactor.name}/url
tagHEAD/tag
/scm


2015-08-11 21:31 GMT+02:00 Robert Scholte (JIRA) j...@apache.org:


 [
 https://issues.apache.org/jira/browse/MPIR-234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14682341#comment-14682341
 ]

 Robert Scholte commented on MPIR-234:
 -

 The more I think about it, the more I like the idea of [~stephenc].
 - No scm specific logic in Maven core.
 - Relying on the scm of the parent isn't always correct. Instead only a
 moduled pom can calculate the actual connection for its modules.
 - For the m-release-p it'll be much easier: no scm on the execution root,
 then it is not a valid release root. This will prevent a lot of issues when
 starting a new Maven project which fail when tagging due to incorrect
 inheritance.
 This is actually one of the things I tried to fix with
 maven-project-utils' [ScmUtils.java|
 http://svn.apache.org/viewvc/maven/shared/trunk/maven-project-utils/src/main/java/org/apache/maven/shared/project/utils/ScmUtils.java?view=markup
 ].
 If we want to expose the connection for submodules like now is done for
 svn, we could adjust the pom.xml during a release: all poms are touched
 anyway and m-release-p already has the required scm dependency so the
 calculation should be simple.

  SCM-link in site of multimodule projects should not append module name
 by default (at least for git)
 
 
 
  Key: MPIR-234
  URL: https://issues.apache.org/jira/browse/MPIR-234
  Project: Maven Project Info Reports Plugin
   Issue Type: Bug
   Components: scm
 Affects Versions: 2.4
 Reporter: Mirko Friedenhagen
 
  I have setup a simple multi module project (see
 https://github.com/mfriedenhagen/multi-module-sample/tree/multi-site-complex)
 which uses git on github as {{scm}}. While rendering the site, MPIR will by
 default add the name of the module to the SCM-URLs in
 source-repository.html. So instead of
 https://github.com/mfriedenhagen/multi-module-sample/ I see
 https://github.com/mfriedenhagen/multi-module-sample/core/,
 g...@github.com:mfriedenhagen/multi-module-sample.git/core and git://
 github.com/mfriedenhagen/multi-module-sample.git/core in the report for
 the core module. All these URLs are invalid. For SVN this could be assumed
 to be the right behaviour, for git and probably other SCMs this is not
 true. As a workaround I have to reconfigure the scm section (see
 https://github.com/mfriedenhagen/multi-module-sample/blob/multi-site-complex/core/pom.xml)
 in the modules like this:
  {code:xml}
  scm
connection${project.parent.scm.connection}/connection
 
  
 

Re: svn commit: r1695142 - /maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

2015-08-13 Thread Lukas Theussl


see related https://issues.apache.org/jira/browse/DOXIA-436

Greetings! :)
-Lukas


Am 12.08.2015 um 23:54 schrieb herve.bout...@free.fr:

now that you told it, I'd seriously prefer change Doxia Markdown parser to use 
an XhtmlParser instance internally instead of extending XhtmlParser while 
completely replacing content parsed by the Xhtml parser: this would be a lot 
more clear (and would avoid adding bloat to getType())

I didn't really try, I don't know if this change is really complex

did you try?

Regards,

Hervé

- Mail original -
De: Petar Tahchiev paranoia...@gmail.com
À: Maven Developers List dev@maven.apache.org
Envoyé: Mercredi 12 Août 2015 08:39:12
Objet: Re: svn commit: r1695142 - 
/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

OK,

sorry I wasn't aware that user specifying an input encoding for xml file
would be considered as introducing a bug. Great for the test-case - I will
revert my changes and work for a fix in the MarkdownParser. Would
overriding the getType() method of the MarkdownParser be considered as a
valid solution?

2015-08-12 2:42 GMT+03:00 herve.bout...@free.fr:


IIUC, your concerns are about Mardown: if Markdown parser has a bug, don't
hesitate to fix it
but do not break content for normal XML parsers, like fml or xdoc

since your change did not make unit tests fail, this proves unit tests are
too weak: I just improved them in r1695408 to fail (and show clearly what
you are breaking)
and I reopened DOXIASITETOOLS-104

You're probably right that making Markdown parser *extend* XhtmlParser is
probably wrong: it should *use* an XhtmlParser, but not extend it

Regards,

Hervé

- Mail original -
De: Petar Tahchiev paranoia...@gmail.com
À: Maven Developers List dev@maven.apache.org
Envoyé: Mardi 11 Août 2015 11:36:28
Objet: Re: svn commit: r1695142 -
/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

Hi Herve,

I did this because seems like a parser can be of type XML even if it's not
parsing only XML - for example the MarkdownParser (which is in doxia and
extends from the XmlParser) getType() returns 2 (XML parser type). I guess
there are two ways to go here - 1) first would be to allow the user to
force an encoding. It's his/hers decision and he/she takes the
responsibility. 2) Would be to override the XmlParser:getType() method in
MarkdownParser and make it return 0 (UNKNOWN_TYPE). To me this would lead
to inconsistency, because the MarkdownParser extends from XmlParser, but
returns another type. Furthermore I don't agree markdown syntax is in fact
xml syntax.


2015-08-11 11:04 GMT+03:00 herve.bout...@free.fr:


wow, I don't like this
in XML, encoding is self provided

with such feature, an XML-invalid document can be read by Maven (and

Maven

only, since it is XML-invalid)

I'm -1 on this: we can't help people make Maven-specific pseudo XML

Regards,

Hervé

- Mail original -
De: ptahch...@apache.org
À: comm...@maven.apache.org
Envoyé: Lundi 10 Août 2015 20:00:00
Objet: svn commit: r1695142 -


/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java


Author: ptahchiev
Date: Mon Aug 10 18:00:00 2015
New Revision: 1695142

URL: http://svn.apache.org/r1695142
Log:
Check for user's provided encoding, and only if it's null then use the
encoding of the xml document. Closes [DOXIASITETOOLS-104]

Modified:



maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java


Modified:


maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

URL:


http://svn.apache.org/viewvc/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java?rev=1695142r1=1695141r2=1695142view=diff




==

---


maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

(original)
+++


maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

Mon Aug 10 18:00:00 2015
@@ -389,7 +389,14 @@ public class DefaultSiteRenderer
  switch ( parser.getType() )
  {
  case Parser.XML_TYPE:
-reader = ReaderFactory.newXmlReader( doc );
+if ( siteContext.getInputEncoding() != null )
+{
+reader = ReaderFactory.newReader( doc,
siteContext.getInputEncoding() );
+}
+else
+{
+reader = 

[jira] [Commented] (MPIR-234) SCM-link in site of multimodule projects should not append module name by default (at least for git)

2015-08-13 Thread Chris Graham (JIRA)

[ 
https://issues.apache.org/jira/browse/MPIR-234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14696261#comment-14696261
 ] 

Chris Graham commented on MPIR-234:
---

it's a general contract that should be enforcer for every supported scm 
(http://maven.apache.org/scm/scms-overview.html ): scm url should support going 
from a parent to a child Maven module by adding the path to the url

One that made perfect sense in the CVS/SVN days, and these core concepts have 
been etched in stone in the Maven  SCM thinking. However, this is not always 
possible with all SCM's. 

The base assumption above, does not always hold true for Accurev, ClearCase, 
RTC, and depending on how you set it up, possibly even Git.

Git can be nested, with a root pom (and moduleblah/module statements), or 
flattened (and module../blah/module statements). If flattened, the 
assumption above does not told true.


 SCM-link in site of multimodule projects should not append module name by 
 default (at least for git)
 

 Key: MPIR-234
 URL: https://issues.apache.org/jira/browse/MPIR-234
 Project: Maven Project Info Reports Plugin
  Issue Type: Bug
  Components: scm
Affects Versions: 2.4
Reporter: Mirko Friedenhagen

 I have setup a simple multi module project (see 
 https://github.com/mfriedenhagen/multi-module-sample/tree/multi-site-complex) 
 which uses git on github as {{scm}}. While rendering the site, MPIR will by 
 default add the name of the module to the SCM-URLs in source-repository.html. 
 So instead of https://github.com/mfriedenhagen/multi-module-sample/ I see 
 https://github.com/mfriedenhagen/multi-module-sample/core/, 
 g...@github.com:mfriedenhagen/multi-module-sample.git/core and 
 git://github.com/mfriedenhagen/multi-module-sample.git/core in the report for 
 the core module. All these URLs are invalid. For SVN this could be assumed to 
 be the right behaviour, for git and probably other SCMs this is not true. As 
 a workaround I have to reconfigure the scm section (see 
 https://github.com/mfriedenhagen/multi-module-sample/blob/multi-site-complex/core/pom.xml)
  in the modules like this:
 {code:xml}
 scm
   connection${project.parent.scm.connection}/connection
   
 developerConnection${project.parent.scm.developerConnection}/developerConnection
   url${project.parent.scm.url}/url
 /scm
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org



[jira] [Commented] (MPIR-234) SCM-link in site of multimodule projects should not append module name by default (at least for git)

2015-08-13 Thread Chris Graham (JIRA)

[ 
https://issues.apache.org/jira/browse/MPIR-234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14696247#comment-14696247
 ] 

Chris Graham commented on MPIR-234:
---

I see a general drift towards a Git centric discussion, which is understandable 
as most of us use it. However, the entire Maven (and particularly the SCM APIs) 
implementation (and therefore thinking) are meant to be SCM agnostic.

As for Flattened/Nested. Let me try to explain (I was rushed). When you perform 
a release with MRP, you perform a tagging operation. With SVN, it's a path 
within the repo. We only tag our own bit. In Git, the entire repo is tagged, 
and in RTC, the Component has a Snapshot taken. In SVN you can check out just a 
path into your WC - you can even cherry pick from where and create your own 
working structure. This falls down as the MRP can not reconstruct this, hand 
made, cherry picked dir structure; it can only check out one point, and thus 
lends itself to a nest structure (if not enforces it). In Git, you clone and 
entire repo, and in RTC/Clearcase you load an entire Component. RTC, like Git, 
can have multiple folders worth (I think of them as eclipse projects - as 
that's the space I work in - esp with RTC) available. In Git, it's easy (and 
makes sense to) have a root pom and nested modules, an a dir format supported 
by the MRP. In RTC, it's impossible to do with the Eclipse client to place a 
pom.xml in the root of a component (although it is not forbidden from the 
underlying SCM impl, it may be possible from the SCM command line). It really 
wants one or more folders (ie eclipse projects) in the root level of a 
Component. When you check them out (SCM load in rtc terms) you get all of the 
projects/folders into the one check out dir. This is kind of forced by the 
eclipse client as it looks for .project files when it checks out. So you end up 
with a flattened structure. In all of the RTC examples I've work with, esp when 
working with the MRP and the RTC SCM integration, I've always nested the 
projects and manually created links in Eclipse. Most RTC based devs either do 
not know a) you can actually do this, or b) do not want too. There is no 
Project Set Support for RTC SCM in eclipse (although it's been requested). Does 
that all make sense or have I actually made things worse?

In terms of ignore scm section that was more a reference to the 
help:effective-pom comment, where the scm section of the pom is output in all 
pom's not just the root one. I thought that the discussion may have been 
heading in the direction of actually using the scm sections in non root poms, 
as I seem to remember that there is/was a request to actually implement using 
the sub module poms (non root ones) scm section, so you'd perform multiple 
check out operations (to make the hand crafted, cherry picked option listed 
above. This is not a thing that I am a fan of.

My point here, is that I don't think that the scm section should exist, or be 
used, or displayed for anything other than the root pom.



 SCM-link in site of multimodule projects should not append module name by 
 default (at least for git)
 

 Key: MPIR-234
 URL: https://issues.apache.org/jira/browse/MPIR-234
 Project: Maven Project Info Reports Plugin
  Issue Type: Bug
  Components: scm
Affects Versions: 2.4
Reporter: Mirko Friedenhagen

 I have setup a simple multi module project (see 
 https://github.com/mfriedenhagen/multi-module-sample/tree/multi-site-complex) 
 which uses git on github as {{scm}}. While rendering the site, MPIR will by 
 default add the name of the module to the SCM-URLs in source-repository.html. 
 So instead of https://github.com/mfriedenhagen/multi-module-sample/ I see 
 https://github.com/mfriedenhagen/multi-module-sample/core/, 
 g...@github.com:mfriedenhagen/multi-module-sample.git/core and 
 git://github.com/mfriedenhagen/multi-module-sample.git/core in the report for 
 the core module. All these URLs are invalid. For SVN this could be assumed to 
 be the right behaviour, for git and probably other SCMs this is not true. As 
 a workaround I have to reconfigure the scm section (see 
 https://github.com/mfriedenhagen/multi-module-sample/blob/multi-site-complex/core/pom.xml)
  in the modules like this:
 {code:xml}
 scm
   connection${project.parent.scm.connection}/connection
   
 developerConnection${project.parent.scm.developerConnection}/developerConnection
   url${project.parent.scm.url}/url
 /scm
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org



Re: svn commit: r1695142 - /maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

2015-08-13 Thread herve . boutemy
Hi Lukas: nice to se that it has catched your interest :)

Yes, DOXIA-436 is the ideal, where we completely remove any Xhtml intermediate

The idea here is less ambitious: just avoid extending XhtmlParser, since the 
source is not Xhtml/Xml

But of course, don't hesitate to help!

Regards,

Hervé

- Mail original -
De: Lukas Theussl ltheu...@gmail.com
À: Maven Developers List dev@maven.apache.org
Envoyé: Jeudi 13 Août 2015 10:30:21
Objet: Re: svn commit: r1695142 - 
/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java


see related https://issues.apache.org/jira/browse/DOXIA-436

Greetings! :)
-Lukas


Am 12.08.2015 um 23:54 schrieb herve.bout...@free.fr:
 now that you told it, I'd seriously prefer change Doxia Markdown parser to 
 use an XhtmlParser instance internally instead of extending XhtmlParser while 
 completely replacing content parsed by the Xhtml parser: this would be a lot 
 more clear (and would avoid adding bloat to getType())

 I didn't really try, I don't know if this change is really complex

 did you try?

 Regards,

 Hervé

 - Mail original -
 De: Petar Tahchiev paranoia...@gmail.com
 À: Maven Developers List dev@maven.apache.org
 Envoyé: Mercredi 12 Août 2015 08:39:12
 Objet: Re: svn commit: r1695142 - 
 /maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

 OK,

 sorry I wasn't aware that user specifying an input encoding for xml file
 would be considered as introducing a bug. Great for the test-case - I will
 revert my changes and work for a fix in the MarkdownParser. Would
 overriding the getType() method of the MarkdownParser be considered as a
 valid solution?

 2015-08-12 2:42 GMT+03:00 herve.bout...@free.fr:

 IIUC, your concerns are about Mardown: if Markdown parser has a bug, don't
 hesitate to fix it
 but do not break content for normal XML parsers, like fml or xdoc

 since your change did not make unit tests fail, this proves unit tests are
 too weak: I just improved them in r1695408 to fail (and show clearly what
 you are breaking)
 and I reopened DOXIASITETOOLS-104

 You're probably right that making Markdown parser *extend* XhtmlParser is
 probably wrong: it should *use* an XhtmlParser, but not extend it

 Regards,

 Hervé

 - Mail original -
 De: Petar Tahchiev paranoia...@gmail.com
 À: Maven Developers List dev@maven.apache.org
 Envoyé: Mardi 11 Août 2015 11:36:28
 Objet: Re: svn commit: r1695142 -
 /maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

 Hi Herve,

 I did this because seems like a parser can be of type XML even if it's not
 parsing only XML - for example the MarkdownParser (which is in doxia and
 extends from the XmlParser) getType() returns 2 (XML parser type). I guess
 there are two ways to go here - 1) first would be to allow the user to
 force an encoding. It's his/hers decision and he/she takes the
 responsibility. 2) Would be to override the XmlParser:getType() method in
 MarkdownParser and make it return 0 (UNKNOWN_TYPE). To me this would lead
 to inconsistency, because the MarkdownParser extends from XmlParser, but
 returns another type. Furthermore I don't agree markdown syntax is in fact
 xml syntax.


 2015-08-11 11:04 GMT+03:00 herve.bout...@free.fr:

 wow, I don't like this
 in XML, encoding is self provided

 with such feature, an XML-invalid document can be read by Maven (and
 Maven
 only, since it is XML-invalid)

 I'm -1 on this: we can't help people make Maven-specific pseudo XML

 Regards,

 Hervé

 - Mail original -
 De: ptahch...@apache.org
 À: comm...@maven.apache.org
 Envoyé: Lundi 10 Août 2015 20:00:00
 Objet: svn commit: r1695142 -

 /maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

 Author: ptahchiev
 Date: Mon Aug 10 18:00:00 2015
 New Revision: 1695142

 URL: http://svn.apache.org/r1695142
 Log:
 Check for user's provided encoding, and only if it's null then use the
 encoding of the xml document. Closes [DOXIASITETOOLS-104]

 Modified:


 maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java

 Modified:

 maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java
 URL:

 http://svn.apache.org/viewvc/maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java?rev=1695142r1=1695141r2=1695142view=diff


 ==
 ---

 maven/doxia/doxia-sitetools/trunk/doxia-site-renderer/src/main/java/org/apache/maven/doxia/siterenderer/DefaultSiteRenderer.java
 (original)
 +++

 

[jira] [Commented] (MPIR-234) SCM-link in site of multimodule projects should not append module name by default (at least for git)

2015-08-13 Thread Chris Graham (JIRA)

[ 
https://issues.apache.org/jira/browse/MPIR-234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14696274#comment-14696274
 ] 

Chris Graham commented on MPIR-234:
---

So the question becomes what do we put on the Source Repository page for a 
module without an explicit scm section.

We don't. Either N/A or See Root SCM URL. 

This has just really driven it home for me: In SCM's like RTC/Clearcase/Accurev 
(and a flattened Git), the concept of a SCM URL for a submodule makes no sense 
at all. It is not possible. When you check out from these, it is an all or 
nothing approach. You can not cherry pick the bits you want.

So, from my SCM agnostic POV, the poms SCM section is only of value (and valid) 
in a a root (release) pom. It should not be displayed, or inferred from any 
other (ie sub module) location.


 SCM-link in site of multimodule projects should not append module name by 
 default (at least for git)
 

 Key: MPIR-234
 URL: https://issues.apache.org/jira/browse/MPIR-234
 Project: Maven Project Info Reports Plugin
  Issue Type: Bug
  Components: scm
Affects Versions: 2.4
Reporter: Mirko Friedenhagen

 I have setup a simple multi module project (see 
 https://github.com/mfriedenhagen/multi-module-sample/tree/multi-site-complex) 
 which uses git on github as {{scm}}. While rendering the site, MPIR will by 
 default add the name of the module to the SCM-URLs in source-repository.html. 
 So instead of https://github.com/mfriedenhagen/multi-module-sample/ I see 
 https://github.com/mfriedenhagen/multi-module-sample/core/, 
 g...@github.com:mfriedenhagen/multi-module-sample.git/core and 
 git://github.com/mfriedenhagen/multi-module-sample.git/core in the report for 
 the core module. All these URLs are invalid. For SVN this could be assumed to 
 be the right behaviour, for git and probably other SCMs this is not true. As 
 a workaround I have to reconfigure the scm section (see 
 https://github.com/mfriedenhagen/multi-module-sample/blob/multi-site-complex/core/pom.xml)
  in the modules like this:
 {code:xml}
 scm
   connection${project.parent.scm.connection}/connection
   
 developerConnection${project.parent.scm.developerConnection}/developerConnection
   url${project.parent.scm.url}/url
 /scm
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org