Menu
▾
▴
gitslave-devs — Basic mailing list for developers, gets bug notifications, etc
You can subscribe to this list here.
| 2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(11) |
Oct
(13) |
Nov
(1) |
Dec
(3) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2011 |
Jan
|
Feb
(20) |
Mar
(24) |
Apr
(1) |
May
|
Jun
(2) |
Jul
(2) |
Aug
(1) |
Sep
(2) |
Oct
|
Nov
|
Dec
(15) |
| 2012 |
Jan
(25) |
Feb
(10) |
Mar
(12) |
Apr
|
May
(3) |
Jun
|
Jul
(2) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(5) |
| 2013 |
Jan
(3) |
Feb
|
Mar
|
Apr
(3) |
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
| 2014 |
Jan
|
Feb
(3) |
Mar
(1) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2015 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
|
From: Mark E. H. <mh...@sa...> - 2015-11-06 15:56:31
|
Hi, We're in the process of evaluating a few different tools for combining multiple repositories, of which git slave is one. We have some concerns, however, that it hasn't had much activity in the last few years, even though there are a number of outstanding issues (notably the fact that gits status doesn't work with git-1.8.5 or later; since we are using 2.6.1 this caused us problems.) Is git slave still actively maintained? Is anyone working on issues, or should be be focusing our evaluations elsewhere? If elsewhere, can someone suggest alternatives? Git submodules seems really painful, and google repo is badly documented for use outside of (or, it seems, within) the Android community. -- ---------------- Mark E. Hamilton Engineering Sciences Center Senior Member of Technical Staff Sandia National Laboratories 505-844-7666 |
|
From: Kedar S. <ke...@ma...> - 2014-03-27 06:43:53
|
Please consider for inclusion. 1. Ensure number continuity is retained across repository changes 2. Ensure all patches are generated in a common output directory 3. Ensure the generated patches have the appropriate prefix that are specified in the git-slave configuration Signed-off-by: Kedar Sovani <ke...@ma...> |
|
From: Baka P. <bak...@us...> - 2014-02-20 02:55:46
|
Updated from old rev 57fc532 (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=shortlog;h=82a813c - Log -------------- commit 82a813c http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=commitdiff;h=82a813c * Switch autoinit to better name autoattach, remove idea from BugsTodo -------------------- Summary of changes: code.git/BugsTodo | 9 --------- code.git/README | 12 ++++++------ code.git/gits | 16 ++++++++-------- code.git/prep_gitscheck | 10 +++++----- 4 files changed, 19 insertions(+), 28 deletions(-) -- gitslave |
|
From: Baka P. <bak...@us...> - 2014-02-19 23:26:57
|
Updated from old rev f5db702 (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=shortlog;h=57fc532 - Log -------------- commit 57fc532 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=commitdiff;h=57fc532 * autoinit will set up .gitslave/.gitignore from current checkout Also, put hints in the synopsis about the special commands supported by gits for --help support -------------------- Summary of changes: code.git/README | 45 +++++++++++++++-- code.git/gits | 151 +++++++++++++++++++++++++++++++++++++++++++++++++++++--- code.git/prep_gitscheck | 7 +++ 3 files changed, 190 insertions(+), 13 deletions(-) -- gitslave |
|
From: Baka P. <bak...@us...> - 2014-02-09 23:20:25
|
Updated from old rev ebac052 (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=shortlog;h=f5db702 - Log -------------- commit f5db702 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=code.git/code.git;a=commitdiff;h=f5db702 * gits remote set-url should support URL substitution Note two URL form (<new> [old]) not supported -------------------- Summary of changes: code.git/gits | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) -- gitslave |
|
From: Daniel D. <dan...@ba...> - 2013-12-15 22:51:29
|
Hello,
I'm trying gitslave following the tutorial[1] and find that “gits
status” is broken on my system:
dad@home:~/tmp/super (master)$ gits status
gits unexpected status output (missing branch): at /home/dad/bin/gits line 2960.
Looking at the code I found 2 issues:
1. It match on english git but does not enforce “LANG=C”
2. The “while” loop on line 2954 remove everything from “$msg” so gits
die line 2960.
I force the language to C by the following patch, for the “while” loop
logic I don't know what to do.
Regards.
#+begin_src diff
diff --git a/gits b/gits
index 5d77016..e667eb3 100755
--- a/gits
+++ b/gits
@@ -190,7 +190,7 @@ sub getcmd($;$)
$cmd = "{ $cmd ;} 2>&1";
print STDERR "Running command: `$cmd`\n" if ($OPTIONS{'verbose'} > 1);
- my $out = `$cmd`;
+ my $out = `export LANG=C; $cmd`;
# strip out progress messages (e.g. "Writing objects: 94% (47/50)")
$out =~ s/^.*\d% .*\r(?!\n)//mg;
if ($OPTIONS{'verbose'} > 2)
#+end_src
Footnotes:
[1] http://gitslave.sourceforge.net/tutorial-basic.html
--
Daniel Dehennin
Récupérer ma clef GPG:
gpg --keyserver pgp.mit.edu --recv-keys 0x7A6FE2DF
|
|
From: Gareth C. <gar...@gm...> - 2013-06-10 02:21:17
|
Hi, I have been playing with gitslave for a little while. It looks like a great tool! I was just wondering - has anyone used gitslave with Jenkins? There is no plugin is there (I haven't found one, but was hoping I missed something)? thanks in advance, Gareth |
|
From: Romain L. <Rom...@ma...> - 2013-04-22 07:52:21
|
Le 20/04/13 01:37, pro...@ba... a écrit :
> In message <516...@ma...>, Romain LEGUAY writes:
>
> I'm a new user of Gits and I try to use it on Windows. I build an .exe
> with PAR::Packer but when I try to populate a project, I have the
> following error:
>
> C:\MyApplication>gits.exe populate
> error: invalid key: 'gits.origin.fromcheckout'
> '{' is not recognized as an internal or external command
> operable program or batch file.
> Could not clone git@myserver:folder/project.git onto Application
>
> For information, I obtain the same result when I launch the perl script
> gits with the command perl (perl gits populate).
>
> I use the last version of Strawberry Perl (64-bit): 5.16.3.1-64bit
>
> Has anyone got this problem?
>
> ----------------------------------------------------------------------
>
> I fix a good part of the problem: I replace: "{ $cmd ; } by "$cmd
>
> I still have the message:
> error: invalid key: 'gits.origin.fromcheckout'
>
> I join the modified gits file.
>
> Hope it'll be usefull,
>
> What version of git do you have installed? What version of windows?
I have git version 1.8.1 (msysgit.1) with Windows 7 Pro x64. And I
launch git (and gits) in Dos prompt (cmd).
>
> What happens if you type, at the Git shell (bash) prompt, the command
> git config gits.origin.fromcheckout
Does nothing on Dos prompt and Git Bash
>
> What happens if you type, at the Git shell (bash) prompt, the command
> { git --version; }
Does not work on Dos prompt, display git version on Git Bash
> What happens if you type, at the Git shell (bash) prompt, the characters
> ^X^V (control-x control-v)
Does not work on Dos prompt, display: "GNU bash, version
3.1.0(1)-release (i686-pc-msys)"
>
> What happens if you type, at the Git shell (bash) prompt, the command
> perl -e 'system("{ git --version ; }");'
Display nothing on Dos prompt, display git version on Git Bash.
> I'm just trying to understand how your system differs from the ones
> where everything works.
>
> -Seth Robertson
Some of my users don't want to learn Unix command (like 'ls' and co) so
I'm trying to bring gitslave on Dos.
Thanks again,
Romain
|
|
From: <pro...@ba...> - 2013-04-19 23:37:21
|
In message <516...@ma...>, Romain LEGUAY writes:
I'm a new user of Gits and I try to use it on Windows. I build an .exe
with PAR::Packer but when I try to populate a project, I have the
following error:
C:\MyApplication>gits.exe populate
error: invalid key: 'gits.origin.fromcheckout'
'{' is not recognized as an internal or external command
operable program or batch file.
Could not clone git@myserver:folder/project.git onto Application
For information, I obtain the same result when I launch the perl script
gits with the command perl (perl gits populate).
I use the last version of Strawberry Perl (64-bit): 5.16.3.1-64bit
Has anyone got this problem?
----------------------------------------------------------------------
I fix a good part of the problem: I replace: "{ $cmd ; } by "$cmd
I still have the message:
error: invalid key: 'gits.origin.fromcheckout'
I join the modified gits file.
Hope it'll be usefull,
What version of git do you have installed? What version of windows?
What happens if you type, at the Git shell (bash) prompt, the command
git config gits.origin.fromcheckout
What happens if you type, at the Git shell (bash) prompt, the command
{ git --version; }
What happens if you type, at the Git shell (bash) prompt, the characters
^X^V (control-x control-v)
What happens if you type, at the Git shell (bash) prompt, the command
perl -e 'system("{ git --version ; }");'
I'm just trying to understand how your system differs from the ones
where everything works.
-Seth Robertson
|
|
From: Romain L. <Rom...@ma...> - 2013-04-18 12:44:33
|
Hello everybody,
I'm a new user of Gits and I try to use it on Windows. I build an .exe
with PAR::Packer but when I try to populate a project, I have the
following error:
C:\MyApplication>gits.exe populate
error: invalid key: 'gits.origin.fromcheckout'
'{' is not recognized as an internal or external command
operable program or batch file.
Could not clone git@myserver:folder/project.git onto Application
For information, I obtain the same result when I launch the perl script
gits with the command perl (perl gits populate).
I use the last version of Strawberry Perl (64-bit): 5.16.3.1-64bit
Has anyone got this problem?
Thanks,
Romain
|
|
From: SourceForge.net <no...@so...> - 2013-01-15 16:27:44
|
Bugs item #3600798, was opened at 2013-01-14 08:29 Message generated for change (Comment added) made by sebdyn You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Sebastian (sebdyn) Assigned to: Nobody/Anonymous (nobody) Summary: Reopen bug #3596870 Initial Comment: Hi, I am sorry to open a new bug, but I can not add a comment to bug #3596870 since it is closed for non-member comments. I want to reopen #3596870 because the current Git revision does not resolve the problem issued in #3596870. ---------------------------------------------------------------------- >Comment By: Sebastian (sebdyn) Date: 2013-01-15 08:27 Message: I was wrong with my last comment: It is still there in gits pull AND gits fetch! Here is the output: > gits pull On: ext: From git/sos/%REPO% * branch pu/vis -> FETCH_HEAD Already up-to-date. On: test: From git/sos/%REPO% * branch pu/vis -> FETCH_HEAD Fast-forward test_sos_import/test_sos_import.pro | 1 + test_sos_import/test_stl/test_read_stl_bin.cpp | 56 ++++++++++++++++++++++++ test_sos_import/test_stl/test_read_stl_txt.cpp | 55 +++++++++++++++++++++++ test_sos_import/test_stl/test_stl.pro | 25 +++++++++++ 4 files changed, 137 insertions(+), 0 deletions(-) create mode 100644 test_sos_import/test_stl/test_read_stl_bin.cpp create mode 100644 test_sos_import/test_stl/test_read_stl_txt.cpp create mode 100644 test_sos_import/test_stl/test_stl.pro On: (sos): From git/%REPO%/sos %REPO% * branch %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% pu/vis %REPO% %REPO% %REPO% %REPO% -> FETCH_HEAD Fast-forward %REPO% src/sos_data/structure/element/topology/etedge.cpp | %REPO% %REPO% 36 +++ %REPO% src/sos_import/stl/stl_reader.cpp %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% | %REPO% 260 ++++++++++++++++++++ %REPO% src/sos_toolbox/statistics/quantile.cpp %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% | %REPO% %REPO% %REPO% 2 +- %REPO% 17 files changed, 915 insertions(+), 56 deletions(-) %REPO%On: (sos): %REPO%On: ext, test: gits fetch gives: On: (sos): From git/%REPO%/sos %REPO% * branch %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% %REPO% pu/vis %REPO% %REPO% %REPO% %REPO% -> FETCH_HEAD %REPO%On: ext, test: From git/sos/%REPO% * branch pu/vis -> FETCH_HEAD ---------------------------------------------------------------------- Comment By: Sebastian (sebdyn) Date: 2013-01-14 09:35 Message: I want to refine the bug report: It seems that 'gits pull' is fixed for standard usage. 'gits fetch' however still reports the erroneous output. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2013-01-14 17:35:11
|
Bugs item #3600798, was opened at 2013-01-14 08:29 Message generated for change (Comment added) made by sebdyn You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Sebastian (sebdyn) Assigned to: Nobody/Anonymous (nobody) Summary: Reopen bug #3596870 Initial Comment: Hi, I am sorry to open a new bug, but I can not add a comment to bug #3596870 since it is closed for non-member comments. I want to reopen #3596870 because the current Git revision does not resolve the problem issued in #3596870. ---------------------------------------------------------------------- >Comment By: Sebastian (sebdyn) Date: 2013-01-14 09:35 Message: I want to refine the bug report: It seems that 'gits pull' is fixed for standard usage. 'gits fetch' however still reports the erroneous output. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2013-01-14 16:29:10
|
Bugs item #3600798, was opened at 2013-01-14 08:29 Message generated for change (Tracker Item Submitted) made by sebdyn You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Sebastian (sebdyn) Assigned to: Nobody/Anonymous (nobody) Summary: Reopen bug #3596870 Initial Comment: Hi, I am sorry to open a new bug, but I can not add a comment to bug #3596870 since it is closed for non-member comments. I want to reopen #3596870 because the current Git revision does not resolve the problem issued in #3596870. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3600798&group_id=353214 |
|
From: Baka P. <bak...@us...> - 2012-12-20 21:59:50
|
Updated from old rev 2b1201e (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=shortlog;h=ebac052 - Log -------------- commit ebac052 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=ebac052 * Dylan Simon's suggested patch is better for git-clean -e processing The git-clean manual page usage didn't suggest --exclude (though it was documented further down) or repeated -e options. Supporting both is superior. -------------------- Summary of changes: gitslave/gits | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) -- gitslave |
|
From: SourceForge.net <no...@so...> - 2012-12-20 21:53:00
|
Bugs item #3549494, was opened at 2012-07-26 18:25 Message generated for change (Comment added) made by bakaproject You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None >Status: Closed Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) >Assigned to: Baka Project (bakaproject) Summary: Add support for gits clean -e Initial Comment: From 1.7.3, git clean supports -e to exclude additional paths from cleaning. It would be nice if gits at least passed this argument through, or better did some sensible processing on it. ---------------------------------------------------------------------- >Comment By: Baka Project (bakaproject) Date: 2012-12-20 13:53 Message: Done in recent commit 2b1201e. Note that it was not immediately obvious what "some sensible processing" might be since the argument was a regular expression (otherwise I might have stripped directory prefixes off). if you have a good regular expersion to parse regular expressions, please let me know... ---------------------------------------------------------------------- Comment By: Dylan Simon (dylex) Date: 2012-07-26 19:13 Message: Here's a simple patch to pass it through (though I don't really understand why clean does its own option processing at all): diff --git a/gits b/gits index 7380a95..81aedc9 100755 --- a/gits +++ b/gits @@ -3238,7 +3238,7 @@ elsif ($ARGV[0] eq 'clean') my $CLEAN_USAGE = "Usage: gits clean -f|-n [-dqxX] [--] [PATH]...\n"; my %COPTIONS; - GetOptions(\%COPTIONS, 'd', 'f', 'q', 'n', 'X', 'x') || die $CLEAN_USAGE; + GetOptions(\%COPTIONS, 'd', 'f|force', 'q|quiet', 'n|dry-run', 'X', 'x', 'e|exclude=s@') || die $CLEAN_USAGE; die "Must set exactly one of -n or -f\n$CLEAN_USAGE" unless (!$COPTIONS{'f'} != !$COPTIONS{'n'}); @@ -3255,6 +3255,7 @@ elsif ($ARGV[0] eq 'clean') $cmd .= " -f" if ($COPTIONS{'f'}); $cmd .= " -X" if ($COPTIONS{'X'}); $cmd .= " -x" if ($COPTIONS{'x'}); + $cmd .= " -e ".quoteit($_) for (@{$COPTIONS{'e'}}); $cmd .= " -- ".quoteit(@ARGV) if ($#ARGV >= 0); my ($ret, $msg) = getcmd($cmd,$slave); ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 |
|
From: Baka P. <bak...@us...> - 2012-12-20 21:49:38
|
Updated from old rev e096b16 (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=shortlog;h=2b1201e - Log -------------- commit 2b1201e http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=2b1201e * Support 'git clean -e'. Also update README from previous commits. commit 6bc459b http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=6bc459b * Remove $USAGE in favor of pod2usage, add --long-help for full man page commit 3697648 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=3697648 * prevent update-remote-url from being noisy by default commit 925634d http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=925634d * Allow --keep-going to keep going even if branch does not exist, also bugs Fix `gits --no-master checkout` from claiming that it ran on the master (even though it did not). -------------------- Summary of changes: gitslave/README | 7 ++-- gitslave/gits | 89 ++++++++++++++++++++++++++++++++++--------------------- gitslave/prep_gitscheck | 9 ++++++ 3 files changed, 68 insertions(+), 37 deletions(-) -- gitslave |
|
From: SourceForge.net <no...@so...> - 2012-12-20 21:17:44
|
Bugs item #3596870, was opened at 2012-12-17 03:21 Message generated for change (Comment added) made by bakaproject You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3596870&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None >Status: Closed Resolution: None Priority: 5 Private: No Submitted By: Sebastian (sebdyn) Assigned to: Nobody/Anonymous (nobody) Summary: "Gits pull" returns an output with hundrets of "%REPO%" Initial Comment: I uses Guts v 2.0.2 on various Linux and Mac systems. Our directory structure is the following: /data/develop/a/git/a/a.git (master repo) /data/develop/a/git/a/b.git (slave "b") /data/develop/a/git/a/c.git (slave "c") The Git-slave repo is then organised as: develop/a develop/a/b develop/a/c Now the problem: When using "gits pull" on various Linux platforms, the output for the master repo is full of "%REPO%" strings. It seems that each white space character (space, tabs, end lines) was replaced by "%REPO%". Interestingly, the option gits --rawout pull does not circumvent this problem, but it gives a clue where the origin lies: It tells me the position of the repos and does so wrong for the master repo: gits --rawout pull central next From /data/develop/%REPO%/a/a * branch next -> FETCH_HEAD Already up-to-date. From /data/develop/a/git/a/%REPO% * branch next -> FETCH_HEAD Already up-to-date. From /data/develop/a/git/a/%REPO% * branch next -> FETCH_HEAD So it seems that Git-Slave finds the wrong position in the path name to be replaced (It takes the first found item, not the last found item in the path name string). ---------------------------------------------------------------------- >Comment By: Baka Project (bakaproject) Date: 2012-12-20 13:17 Message: This appears to have been fixed in commit 0fc40b "fix inappropriate %REPO% substitution" in a post 2.0.2 release commit. Please try downloading the latest git version of gitslave and see if that helps. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3596870&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-12-17 11:21:31
|
Bugs item #3596870, was opened at 2012-12-17 03:21 Message generated for change (Tracker Item Submitted) made by sebdyn You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3596870&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Sebastian (sebdyn) Assigned to: Nobody/Anonymous (nobody) Summary: "Gits pull" returns an output with hundrets of "%REPO%" Initial Comment: I uses Guts v 2.0.2 on various Linux and Mac systems. Our directory structure is the following: /data/develop/a/git/a/a.git (master repo) /data/develop/a/git/a/b.git (slave "b") /data/develop/a/git/a/c.git (slave "c") The Git-slave repo is then organised as: develop/a develop/a/b develop/a/c Now the problem: When using "gits pull" on various Linux platforms, the output for the master repo is full of "%REPO%" strings. It seems that each white space character (space, tabs, end lines) was replaced by "%REPO%". Interestingly, the option gits --rawout pull does not circumvent this problem, but it gives a clue where the origin lies: It tells me the position of the repos and does so wrong for the master repo: gits --rawout pull central next From /data/develop/%REPO%/a/a * branch next -> FETCH_HEAD Already up-to-date. From /data/develop/a/git/a/%REPO% * branch next -> FETCH_HEAD Already up-to-date. From /data/develop/a/git/a/%REPO% * branch next -> FETCH_HEAD So it seems that Git-Slave finds the wrong position in the path name to be replaced (It takes the first found item, not the last found item in the path name string). ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3596870&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-07-27 02:13:23
|
Bugs item #3549494, was opened at 2012-07-26 18:25 Message generated for change (Comment added) made by dylex You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Add support for gits clean -e Initial Comment: From 1.7.3, git clean supports -e to exclude additional paths from cleaning. It would be nice if gits at least passed this argument through, or better did some sensible processing on it. ---------------------------------------------------------------------- Comment By: Dylan Simon (dylex) Date: 2012-07-26 19:13 Message: Here's a simple patch to pass it through (though I don't really understand why clean does its own option processing at all): diff --git a/gits b/gits index 7380a95..81aedc9 100755 --- a/gits +++ b/gits @@ -3238,7 +3238,7 @@ elsif ($ARGV[0] eq 'clean') my $CLEAN_USAGE = "Usage: gits clean -f|-n [-dqxX] [--] [PATH]...\n"; my %COPTIONS; - GetOptions(\%COPTIONS, 'd', 'f', 'q', 'n', 'X', 'x') || die $CLEAN_USAGE; + GetOptions(\%COPTIONS, 'd', 'f|force', 'q|quiet', 'n|dry-run', 'X', 'x', 'e|exclude=s@') || die $CLEAN_USAGE; die "Must set exactly one of -n or -f\n$CLEAN_USAGE" unless (!$COPTIONS{'f'} != !$COPTIONS{'n'}); @@ -3255,6 +3255,7 @@ elsif ($ARGV[0] eq 'clean') $cmd .= " -f" if ($COPTIONS{'f'}); $cmd .= " -X" if ($COPTIONS{'X'}); $cmd .= " -x" if ($COPTIONS{'x'}); + $cmd .= " -e ".quoteit($_) for (@{$COPTIONS{'e'}}); $cmd .= " -- ".quoteit(@ARGV) if ($#ARGV >= 0); my ($ret, $msg) = getcmd($cmd,$slave); ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-07-27 01:25:22
|
Bugs item #3549494, was opened at 2012-07-26 18:25 Message generated for change (Tracker Item Submitted) made by nobody You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Add support for gits clean -e Initial Comment: From 1.7.3, git clean supports -e to exclude additional paths from cleaning. It would be nice if gits at least passed this argument through, or better did some sensible processing on it. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3549494&group_id=353214 |
|
From: Baka P. <bak...@us...> - 2012-05-08 15:42:38
|
Updated from old rev 49fea09 (commit) http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=shortlog;h=e096b16 - Log -------------- commit e096b16 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=e096b16 * gits-enable is entirely useless since gits-attach fails when used commit 60c852d http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=60c852d * Support attaching subslaves to slaves which you do not want to change Inspired by message on git mailing list commit 8cb02f0 http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=commitdiff;h=8cb02f0 * reorder likely candidates up top -------------------- Summary of changes: gitslave/BugsTodo | 308 ++++++++++++++++++++++++++++++++++++-------------------------- 1 files changed, 181 insertions(+), 127 deletions(-) -- gitslave |
|
From: SourceForge.net <no...@so...> - 2012-05-08 15:41:06
|
Bugs item #3523540, was opened at 2012-05-04 08:31 Message generated for change (Comment added) made by bakaproject You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3523540&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None >Status: Closed >Resolution: Works For Me Priority: 5 Private: No Submitted By: Mark (mmindenhall) >Assigned to: Baka Project (bakaproject) Summary: Feature request: track slave without cloning under master Initial Comment: The "gits attach" command requires a REPOSITORY argument (path to repo already cloned locally), and a LOCALPATH argument (under the master repo, where the "slave" repo will be cloned again). The end result is a second set of slave repo folders under the master repo. This makes sense when the slaves are submodules or subprojects of the master. However, I'm working with several related but independent repos, where there's no clear project/subproject relationship, and where I need to be able to operate across them simultaneously (e.g., create branches, push branches to origin, etc.). What's more, some of these are quite large repos (containing a lot of image content as well as code), and I'd prefer not to have duplicate copies of the repos on my disk. I'm therefore wondering if it would be possible for gits to track repos as slaves wherever they've already been cloned, without cloning them again under the master repo. I tried passing the same path for the REPOSITORY and LOCALPATH args of "gits attach", but I get the message "Destination (<folder>) already exists". I hope that's clear...basically I just want to be able to designate any git repo as the master, and any other git repo as a slave, regardless of where they exist on my hard drive, and without cloning the slaves into folders under the master repo. I realize this would break the gits ability for a repo to be a slave in multiple masters (e.g., library used by multiple projects), but that's fine for my use case. ---------------------------------------------------------------------- >Comment By: Baka Project (bakaproject) Date: 2012-05-08 08:41 Message: If I understand your problem, then gitslave already has an option to handle this. In the 2.0.2 release, we added the "--adminonly" flag to gits-attach which only makes the entry in .gitslave and does not attempt to clone. Since you have already cloned, this would seem to be ideal. Gitslave has not been really tested in an environment where the slave repos are not underneath the master directory/repo, but it didn't obviously fail in a quick test. Obviously all gitslave users would have to have the same on-disk repository layout on their local systems. Also note that gitslave doesn't currently functionally support management of .gitslave files which are not committed into git. It isn't clear for your use case whether this matters or not, but if it does, you may need to stick the right bits in the right files. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3523540&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-05-04 15:31:31
|
Bugs item #3523540, was opened at 2012-05-04 08:31 Message generated for change (Tracker Item Submitted) made by mmindenhall You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3523540&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: core Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Mark (mmindenhall) Assigned to: Nobody/Anonymous (nobody) Summary: Feature request: track slave without cloning under master Initial Comment: The "gits attach" command requires a REPOSITORY argument (path to repo already cloned locally), and a LOCALPATH argument (under the master repo, where the "slave" repo will be cloned again). The end result is a second set of slave repo folders under the master repo. This makes sense when the slaves are submodules or subprojects of the master. However, I'm working with several related but independent repos, where there's no clear project/subproject relationship, and where I need to be able to operate across them simultaneously (e.g., create branches, push branches to origin, etc.). What's more, some of these are quite large repos (containing a lot of image content as well as code), and I'd prefer not to have duplicate copies of the repos on my disk. I'm therefore wondering if it would be possible for gits to track repos as slaves wherever they've already been cloned, without cloning them again under the master repo. I tried passing the same path for the REPOSITORY and LOCALPATH args of "gits attach", but I get the message "Destination (<folder>) already exists". I hope that's clear...basically I just want to be able to designate any git repo as the master, and any other git repo as a slave, regardless of where they exist on my hard drive, and without cloning the slaves into folders under the master repo. I realize this would break the gits ability for a repo to be a slave in multiple masters (e.g., library used by multiple projects), but that's fine for my use case. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3523540&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-03-02 17:17:25
|
Bugs item #3495770, was opened at 2012-02-29 10:49 Message generated for change (Comment added) made by dupuy You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3495770&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. >Category: core Group: None >Status: Closed >Resolution: Fixed Priority: 5 Private: No Submitted By: Doug Beatty (dt25954) >Assigned to: Baka Project (bakaproject) Summary: Unable to gits pull or gits fetch Initial Comment: I am having an issue using gits to pull or fetch. All of the environment details are the same as the issue I entered yesterday. Same .gitslave file using Msysgit, etc. I can provide more detail if necessary. See attached screen print for commands and results. ---------------------------------------------------------------------- >Comment By: Alexander Dupuy (dupuy) Date: 2012-03-02 09:17 Message: The use of %REPO% is intentional; in the output from each git command in slave repositories, the repository name is replaced with %REPO%. This allows otherwise identical messages from multiple slave repositories to be combined and displayed only once. If your gitslave master project is combining repositories from multiple unrelated projects, this isn't likely to be useful, but in the project for which we first developed gitslave, we had literally dozens of sub-projects that we were developing, and things like tags and branches were frequently applied to all of them via `gits tag` - by eliminating these trivial differences (repo name, commit hashes, and some other insignificant variations, we were able to combine the messages from these dozens of slave repos and display them just once, which made the output from gits much easier for humans to understand. If this really bothers you, you can use --rawout to get the raw output of gits (you can also use --no-hide to just disable the hash suppression, not %REPO% substitution). ---------------------------------------------------------------------- Comment By: Doug Beatty (dt25954) Date: 2012-03-01 14:23 Message: After looking at the message again it looks like maybe that entire line should not be there? This line: From wdc-scmapp01.m.dstcorp.net:VisionPOC/sp0071143/%REPO% ---------------------------------------------------------------------- Comment By: Doug Beatty (dt25954) Date: 2012-03-01 14:21 Message: No worries. I like this script and your response time on my issues has been excellent. I have pulled the latest code, and verified that I can fetch and pull. It appears to work. However, there is one oddity in the message on the gits fetch. It has %REPO% at the end of the base url. Example follows: dt25954@DCW00149732XP /c/WorkAreas/Vision (master) $ gits fetch sp0071143 On: (Vision), Vision_Core, Vision_Java_Source, Vision_Web, Vision_Static_Content From wdc-scmapp01.m.dstcorp.net:VisionPOC/sp0071143/%REPO% * [new branch] built -> sp0071143/built * [new branch] master -> sp0071143/master ---------------------------------------------------------------------- Comment By: Baka Project (bakaproject) Date: 2012-03-01 12:09 Message: Sorry about that. The problem should be fixed with the latest push to sourceforge. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3495770&group_id=353214 |
|
From: SourceForge.net <no...@so...> - 2012-03-01 22:23:43
|
Bugs item #3495770, was opened at 2012-02-29 10:49 Message generated for change (Comment added) made by dt25954 You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3495770&group_id=353214 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Doug Beatty (dt25954) Assigned to: Nobody/Anonymous (nobody) Summary: Unable to gits pull or gits fetch Initial Comment: I am having an issue using gits to pull or fetch. All of the environment details are the same as the issue I entered yesterday. Same .gitslave file using Msysgit, etc. I can provide more detail if necessary. See attached screen print for commands and results. ---------------------------------------------------------------------- Comment By: Doug Beatty (dt25954) Date: 2012-03-01 14:23 Message: After looking at the message again it looks like maybe that entire line should not be there? This line: From wdc-scmapp01.m.dstcorp.net:VisionPOC/sp0071143/%REPO% ---------------------------------------------------------------------- Comment By: Doug Beatty (dt25954) Date: 2012-03-01 14:21 Message: No worries. I like this script and your response time on my issues has been excellent. I have pulled the latest code, and verified that I can fetch and pull. It appears to work. However, there is one oddity in the message on the gits fetch. It has %REPO% at the end of the base url. Example follows: dt25954@DCW00149732XP /c/WorkAreas/Vision (master) $ gits fetch sp0071143 On: (Vision), Vision_Core, Vision_Java_Source, Vision_Web, Vision_Static_Content From wdc-scmapp01.m.dstcorp.net:VisionPOC/sp0071143/%REPO% * [new branch] built -> sp0071143/built * [new branch] master -> sp0071143/master ---------------------------------------------------------------------- Comment By: Baka Project (bakaproject) Date: 2012-03-01 12:09 Message: Sorry about that. The problem should be fixed with the latest push to sourceforge. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=1475673&aid=3495770&group_id=353214 |
