Quantcast

Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

classic Classic list List threaded Threaded
23 messages Options
12
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Scot Forbes

I could not find a way to tell the SCM Activity plugin to ignore whitespace changes with Git. In effect using the –w flag with the git blame command.

 

Without this ability I get no commit info in the source code viewer due to a known issue using core.autocrlf = true with msysgit on Windows, combined with the fact that the plugin is working against the local copy and not a branch:

 

http://stackoverflow.com/questions/4638500/git-blame-showing-no-history

 

Has anyone come up with a workaround for this?

 

-Scot

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

David Gageot
Hi,

Unfortunately, there's no way to tell SCM Activity Plugin to use -w flag.
We depend on Maven Scm Plugin (http://maven.apache.org/scm/maven-scm-plugin/) and it doesn't know how to do that. 

I've created two tickets: one our our side and one on their side:

David.

On 22 June 2012 02:06, Scot Forbes <[hidden email]> wrote:

I could not find a way to tell the SCM Activity plugin to ignore whitespace changes with Git. In effect using the –w flag with the git blame command.

 

Without this ability I get no commit info in the source code viewer due to a known issue using core.autocrlf = true with msysgit on Windows, combined with the fact that the plugin is working against the local copy and not a branch:

 

http://stackoverflow.com/questions/4638500/git-blame-showing-no-history

 

Has anyone come up with a workaround for this?

 

-Scot




--
-- David
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Scot Forbes

Thank you David!

 

Another workaround would be to allow us to optionally specify the branch to the git blame command instead of using the working copy. Is this also a limitation of the Mavin plugin?

 

Thanks again,

-Scot

 

From: David Gageot [mailto:[hidden email]]
Sent: Monday, June 25, 2012 12:37 AM
To: [hidden email]
Subject: Re: [sonar-user] Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

 

Hi,

 

Unfortunately, there's no way to tell SCM Activity Plugin to use -w flag.

We depend on Maven Scm Plugin (http://maven.apache.org/scm/maven-scm-plugin/) and it doesn't know how to do that. 

 

I've created two tickets: one our our side and one on their side:

 

David.

On 22 June 2012 02:06, Scot Forbes <[hidden email]> wrote:

I could not find a way to tell the SCM Activity plugin to ignore whitespace changes with Git. In effect using the –w flag with the git blame command.

 

Without this ability I get no commit info in the source code viewer due to a known issue using core.autocrlf = true with msysgit on Windows, combined with the fact that the plugin is working against the local copy and not a branch:

 

http://stackoverflow.com/questions/4638500/git-blame-showing-no-history

 

Has anyone come up with a workaround for this?

 

-Scot



 

--
-- David

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

David Gageot
This would be incorrect since sonar would:
 + store blame information for committed copy
 + but analyse violations from local copy

David.

On 25 June 2012 17:53, Scot Forbes <[hidden email]> wrote:

Thank you David!

 

Another workaround would be to allow us to optionally specify the branch to the git blame command instead of using the working copy. Is this also a limitation of the Mavin plugin?

 

Thanks again,

-Scot

 

From: David Gageot [mailto:[hidden email]]
Sent: Monday, June 25, 2012 12:37 AM
To: [hidden email]
Subject: Re: [sonar-user] Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

 

Hi,

 

Unfortunately, there's no way to tell SCM Activity Plugin to use -w flag.

We depend on Maven Scm Plugin (http://maven.apache.org/scm/maven-scm-plugin/) and it doesn't know how to do that. 

 

I've created two tickets: one our our side and one on their side:

 

David.

On 22 June 2012 02:06, Scot Forbes <[hidden email]> wrote:

I could not find a way to tell the SCM Activity plugin to ignore whitespace changes with Git. In effect using the –w flag with the git blame command.

 

Without this ability I get no commit info in the source code viewer due to a known issue using core.autocrlf = true with msysgit on Windows, combined with the fact that the plugin is working against the local copy and not a branch:

 

http://stackoverflow.com/questions/4638500/git-blame-showing-no-history

 

Has anyone come up with a workaround for this?

 

-Scot



 

--
-- David




--
-- David
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Scot Forbes

In my case we are always synchronizing with the committed copy before doing the analysis so I think it would still work for my situation. However I can see how this would not make a good generic solution.

 

-Scot

 

From: David Gageot [mailto:[hidden email]]
Sent: Monday, June 25, 2012 8:57 AM
To: [hidden email]
Subject: Re: [sonar-user] Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

 

This would be incorrect since sonar would:

 + store blame information for committed copy

 + but analyse violations from local copy

 

David.

On 25 June 2012 17:53, Scot Forbes <[hidden email]> wrote:

Thank you David!

 

Another workaround would be to allow us to optionally specify the branch to the git blame command instead of using the working copy. Is this also a limitation of the Mavin plugin?

 

Thanks again,

-Scot

 

From: David Gageot [mailto:[hidden email]]
Sent: Monday, June 25, 2012 12:37 AM
To: [hidden email]
Subject: Re: [sonar-user] Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

 

Hi,

 

Unfortunately, there's no way to tell SCM Activity Plugin to use -w flag.

We depend on Maven Scm Plugin (http://maven.apache.org/scm/maven-scm-plugin/) and it doesn't know how to do that. 

 

I've created two tickets: one our our side and one on their side:

 

David.

On 22 June 2012 02:06, Scot Forbes <[hidden email]> wrote:

I could not find a way to tell the SCM Activity plugin to ignore whitespace changes with Git. In effect using the –w flag with the git blame command.

 

Without this ability I get no commit info in the source code viewer due to a known issue using core.autocrlf = true with msysgit on Windows, combined with the fact that the plugin is working against the local copy and not a branch:

 

http://stackoverflow.com/questions/4638500/git-blame-showing-no-history

 

Has anyone come up with a workaround for this?

 

-Scot



 

--
-- David



 

--
-- David

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa
Hi,

Im facing the same problem, having not commited yet in some msbuild auto generated files (dont ask me why!!).

So it seems on Friday we will have a new release of SCM activity plugin with this fix, excellent :).

But my problem is actually because of this issue the "Not Commited Yet" will not go away even if fixing this problem. The solution i normally use in these cases is to run a new analysis without scm and than again with scm. The problem with this is that my jobs take loads of time (c++ plugin, with loads of external analysis tools) and so i will need at least 2 days of no commit history for developers. Is there a better way of managing these cases in the plugin itself.

Thanks in advance



Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Freddy Mallet
Hi Jorge, 

If you get some "Not Commited Yet" that means that you're launching Sonar analysis locally on a change source code that has not been yet committed/pushed ? 
-----
Sonar for Continuous Inspection



On Wed, Oct 17, 2012 at 8:43 AM, jmecosta <[hidden email]> wrote:
Hi,

Im facing the same problem, having not commited yet in some msbuild auto
generated files (dont ask me why!!).

So it seems on Friday we will have a new release of SCM activity plugin with
this fix, excellent :).

But my problem is actually because of this issue the "Not Commited Yet" will
not go away even if fixing this problem. The solution i normally use in
these cases is to run a new analysis without scm and than again with scm.
The problem with this is that my jobs take loads of time (c++ plugin, with
loads of external analysis tools) and so i will need at least 2 days of no
commit history for developers. Is there a better way of managing these cases
in the plugin itself.

Thanks in advance







-----
Best Regards
Jorge Costa
--
View this message in context: http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004186.html
Sent from the Sonar user mailing list archive at Nabble.com.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email



Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa
Hi Freddy,

this is caused by the setting autocrlf in git. Some of the files in the repository can only seen history if i play with this setting.

In our case, before analysis we do a git reset --hard followed by a git pull to ensure the sources are inline with master repository at that point. After this it uses the clone in the analysis folder.

I think also is that if using the git blame -w on any file than this always works... So im a bit disappointed to see  SONARPLUGINS-1995 not being fixes.

Thanks



Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Julien HENRY-3
Hi Jorge,

Here is what I did as an attempt to reproduce the issue:
  1) create a git repo on Googlecode
  2) populate it from Linux in order to have files with Unix line ends
  3) set up a Windows VM with git for Windows + core.autocrlf=true
  4) clone the repo and check that line ends are actually transformed to Windows CRLF
  5) play with git blame and sonar scm activity

=> I had no problem

So if you can tell me how to be in a similar situation as the one you encounter it will definitely help.

Thanks

Julien


De : jmecosta <[hidden email]>
À : [hidden email]
Envoyé le : Jeudi 18 octobre 2012 8h41
Objet : [sonar-user] RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Hi Freddy,

this is caused by the setting autocrlf in git. Some of the files in the
repository can only seen history if i play with this setting.

In our case, before analysis we do a git reset --hard followed by a git pull
to ensure the sources are inline with master repository at that point. After
this it uses the clone in the analysis folder.

I think also is that if using the git blame -w on any file than this always
works... So im a bit disappointed to see  SONARPLUGINS-1995 not being fixes.

Thanks







-----
Best Regards
Jorge Costa
--
View this message in context: http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004271.html
Sent from the Sonar user mailing list archive at Nabble.com.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email




Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa
This post was updated on .
I can try to create a use case, however one thing we are doing different is the are not using Linux at all, only windows. Repos are created using mygit with git init, and than imported into gitstack. So we have a web interface from where to clone. However the blame commands are always run against current clone, so there are no changes committed in those cases.

This is the procedure:
1. git init .
2. import into gitstack
3. clone from git stack in to a temp clone
4. fetch sources from surround scm
5. commit changes temp clone
6. push changes into gitstack main repository.

On the analysis machine:
1. git clone from gitstack
2. run git blame and than we have problems.

I understand that if it cant be tested there is no point of implement it. so i will try to provide a use case hopefully.



Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Julien HENRY-3
Hi Jorge,

Don't spend time on providing a reproducible test case as I think I have one. The problem occurs with a .Net project hosted on GitHub. I will investigate.

Regards,

Julien


De : jmecosta <[hidden email]>
À : [hidden email]
Envoyé le : Jeudi 18 octobre 2012 11h12
Objet : [sonar-user] RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

I can try to create a use case, however one thing we are doing different is
the are not using Linux at all, only windows. Repos are created using mygit
with git init, and than imported into gitstack. So we have a web interface
from where to clone. However the blame commands are always run against
current clone, so there are no changes committed in those cases.

This is the procedure:
1. git init .
2. import into gitstack
3. clone from git stack in to a temp clone
4. fetch sources from surround scm
5. commit changes temp clone
6. push changes into gitstack main repository.

On the analysis machine:
1. git clone from gitstack
2. run git blame and than we have problems.

I understand that i can be tested there is no point of implement. so i will
try to provide a use case hopefully.







-----
Best Regards
Jorge Costa
--
View this message in context: http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004294.html
Sent from the Sonar user mailing list archive at Nabble.com.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email




Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa
Ok great...

and i confirm that i see loads of those ones also from dot net projects.

Thanks
Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Julien HENRY-3
Jorge,

I finally managed to understand the issue. You were right that it has nothing to do with a mix of Linux/Windows developers.

First I had to perfectly understand what autocrlf=true means. It is clearly explained here [1]. To sum up, it means you want to enforce LF in your repository, but you also want to work with CRLF in your local copy.
When a file is present in your repository with LF and you do a checkout, git will automatically convert the file to CRLF. And when you will commit git will revert CRLF to LF so only your real modifications are sent.

The problematic case is when you have a file that is already stored with CRLF in the repo (maybe the repo was created before setting autocrlf=true). When you do a checkout the file remains with CRLF line ends.
But if you do any modification and commit, git would automatically convert every line ends to LF. And I think this is what git blame is trying to say when it show every line as "Not Committed Yet".

In my personal point of view this is not very consistent with other git commands. For example git status on the same file would not report any pending modification. But it is out of the scope of Sonar SCM Activity plugin.

Considering:
  - that -w flag is not an option because it will produce unwanted side effects
  - that we advocate to run Sonar analysis on a separate infrastructure (= a separate git clone than the one used for development)
  - that Sonar analysis is supposed to run on an accurate copy of source code present in the repository

we will now highly recommend to set core.autocrlf=false in the git configuration of the account responsible for running Sonar analysis. Documentation and JIRA ticket will be updated accordingly.

Please note that this decision still allow you to work with autocrlf=true in your development environment even if autocrlf=true seems to not be very popular these days [2].

Regards,

Julien

[1] http://stackoverflow.com/questions/4181870/git-on-windows-what-do-the-crlf-settings-mean
[2] http://stackoverflow.com/questions/2016673/definitive-recommendation-for-git-autocrlf-settings



De : jmecosta <[hidden email]>
À : [hidden email]
Envoyé le : Jeudi 18 octobre 2012 12h57
Objet : [sonar-user] RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Ok great...

and i confirm that i see loads of those ones also from dot net projects.

Thanks



-----
Best Regards
Jorge Costa
--
View this message in context: http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004305.html
Sent from the Sonar user mailing list archive at Nabble.com.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email




Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa
Hi Julien,

I understand your explanation, so what i did was to re clone everything with same settings in all the clones. And at least now the blame commands are working for all files. However i add to commit a fix to the line endings to get this to work.

So now back to my original question, currently source in sonar still shows not committed yet. I need to regenerate a analyses with scm disable and scm enable. However i will have all the files with same person build blamed. So my question is it possible to only to regenerate the blame in certain files?

Thanks in advance...
Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Julien HENRY-3
Hi Jorge,

Do you mean forcing Sonar to "forget" previous blame result only for certain files? Sonar compute a checksum of each file and will not redo a blame if the checksum is the same (= file not modified since previous analysis).
Line ends are normalized so changing LF to CRLF or the opposite will not trigger a new blame.

In your situation I don't see a simple way to force a refresh of the blame for every files (and also for only a subset).

Several options:
  - delete projet in Sonar and run a new analysis. But project history will be lost!
  - keep project "as is" and wait for files to be modified. If you have files that are never modified it may not be a valid option.
  - create a JIRA ticket on SCM Activity plugin in order to add a parameter that will skip the checksum comparison and consequently do a full blame on every files
  - in the git local repo (= clone) used to perform the sonar analysis, you can try the following tweak to bypass checksum.
      1) Do a dummy modification on every file you want the blame to occurs (for example add a new empty line at the end of file)
      2) Commit (but don't push to not mess main repo)
      3) Perform the Sonar analysis. Checksum will be different so blame will be done.
      4) Revert your commit or delete the repo and redo the clone.

Hope this helps.

Regards,

Julien


----- Mail original -----

> De : jmecosta <[hidden email]>
> À : [hidden email]
> Cc :
> Envoyé le : Vendredi 19 octobre 2012 9h42
> Objet : [sonar-user] RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows
>
> Hi Julien,
>
> I understand your explanation, so what i did was to re clone everything with
> same settings in all the clones. And at least now the blame commands are
> working for all files. However i add to commit a fix to the line endings to
> get this to work.
>
> So now back to my original question, currently source in sonar still shows
> not committed yet. I need to regenerate a analyses with scm disable and scm
> enable. However i will have all the files with same person build blamed. So
> my question is it possible to only to regenerate the blame in certain files?
>
> Thanks in advance...
>
>
>
>
> -----
> Best Regards
> Jorge Costa
> --
> View this message in context:
> http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004365.html
> Sent from the Sonar user mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe from this list, please visit:
>
>     http://xircles.codehaus.org/manage_email
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Julien HENRY-3
As a complement maybe the simplest way is to run an new Sonar analysis with SCM activity plugin disabled (sonar.scm.enabled=false) this way all checksums will be "set" to null.
Then start a new analysis with SCM activity enabled and it will do a blame on every file.

I don't really understand why you would like to do the blame on certain file only?




----- Mail original -----

> De : Julien HENRY <[hidden email]>
> À : "[hidden email]" <[hidden email]>
> Cc :
> Envoyé le : Vendredi 19 octobre 2012 11h42
> Objet : Re: [sonar-user] RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows
>
> Hi Jorge,
>
> Do you mean forcing Sonar to "forget" previous blame result only for
> certain files? Sonar compute a checksum of each file and will not redo a blame
> if the checksum is the same (= file not modified since previous analysis).
> Line ends are normalized so changing LF to CRLF or the opposite will not trigger
> a new blame.
>
> In your situation I don't see a simple way to force a refresh of the blame
> for every files (and also for only a subset).
>
> Several options:
>   - delete projet in Sonar and run a new analysis. But project history will be
> lost!
>   - keep project "as is" and wait for files to be modified. If you
> have files that are never modified it may not be a valid option.
>   - create a JIRA ticket on SCM Activity plugin in order to add a parameter that
> will skip the checksum comparison and consequently do a full blame on every
> files
>   - in the git local repo (= clone) used to perform the sonar analysis, you can
> try the following tweak to bypass checksum.
>       1) Do a dummy modification on every file you want the blame to occurs (for
> example add a new empty line at the end of file)
>       2) Commit (but don't push to not mess main repo)
>       3) Perform the Sonar analysis. Checksum will be different so blame will be
> done.
>       4) Revert your commit or delete the repo and redo the clone.
>
> Hope this helps.
>
> Regards,
>
> Julien
>
>
> ----- Mail original -----
>>  De : jmecosta <[hidden email]>
>>  À : [hidden email]
>>  Cc :
>>  Envoyé le : Vendredi 19 octobre 2012 9h42
>>  Objet : [sonar-user] RE: Problem with SCM Activity plugin and Git using
> core.autocrlf=true on Windows
>>
>>  Hi Julien,
>>
>>  I understand your explanation, so what i did was to re clone everything
> with
>>  same settings in all the clones. And at least now the blame commands are
>>  working for all files. However i add to commit a fix to the line endings to
>>  get this to work.
>>
>>  So now back to my original question, currently source in sonar still shows
>>  not committed yet. I need to regenerate a analyses with scm disable and scm
>>  enable. However i will have all the files with same person build blamed. So
>>  my question is it possible to only to regenerate the blame in certain
> files?
>>
>>  Thanks in advance...
>>
>>
>>
>>
>>  -----
>>  Best Regards
>>  Jorge Costa
>>  --
>>  View this message in context:
>>
> http://sonar.15.n6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5004365.html
>>  Sent from the Sonar user mailing list archive at Nabble.com.
>>
>>  ---------------------------------------------------------------------
>>  To unsubscribe from this list, please visit:
>>
>>      http://xircles.codehaus.org/manage_email
>>
>
> ---------------------------------------------------------------------
> To unsubscribe from this list, please visit:
>
>     http://xircles.codehaus.org/manage_email
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa

Ive normalized the line endings on all the files, so there was a huge commit from a bot. So now all the blame information is pointed to the bot.  Really wanted to reset this info in only a few files because of that.
Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

radu.pop
Hy I have the same problem i put enable to false build and the enable to true and build and no commit yet was there.
The problem is that their are user that are sown for some files and same users are not shown for other files.

Do you have any ideas?
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

Jorge Costa

To debug this you need to have the blame info from the command line. Once its done you can run the analysis in that repository after the false and true set. You might need to fix the line endings to get the blame inform.

The -w option can be use help debug

On Jul 23, 2013 5:56 PM, "radu.pop" <[hidden email]> wrote:
Hy I have the same problem i put enable to false build and the enable to true
and build and no commit yet was there.
The problem is that their are user that are sown for some files and same
users are not shown for other files.

Do you have any ideas?



--
View this message in context: http://sonar.15.x6.nabble.com/Problem-with-SCM-Activity-plugin-and-Git-using-core-autocrlf-true-on-Windows-tp5000095p5014993.html
Sent from the Sonar user mailing list archive at Nabble.com.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Best Regards
Jorge Costa
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Problem with SCM Activity plugin and Git using core.autocrlf=true on Windows

radu.pop
Hi i made the debug as you said.
I'm using git as well and on git blame -w filename

I got on every line some author.
What I identified is that on sonar you get the error on line 41 and you go the blame from line 38. This is good works fine but  on line from 38 to 40 the user from git is inactive/or deleted and on line 41 I have a user that exist and is not inactive. The problem is that for all this lines in sonar interface I get not committed yet.

Here is a kind of log:
what I get with git blame -w:
^a8d011b (Rob Kitson      2012-01-12 08:10:51 -0800  38)
^a8d011b (Rob Kitson      2012-01-12 08:10:51 -0800  39)
^a8d011b (Rob Kitson      2012-01-12 08:10:51 -0800  40)
29f12199 (Radu Pop        2013-07-18 17:36:47 +0300  41)

This is what i get from sonar:

2013-07-24 not.committed.yet 38
                                                39
                                                40
                                                41 <- here should be a name

Hm this is an issue from sonar? Can you please me know if this happens because of sonar of is an internal issue  I'm looking through sonar code to identify this also.
Thanks
12
Loading...