This function should be called to test for /dev/null when nul on Windows can
also be accepted. The is_dev_null function in builtin-apply.c was renamed.
Signed-off-by: Mike Pape <dotzenlabs@gmail.com>
The common way to specify an ssh remote is to say "host:/path",
but MinGW decides for us that this is probably a path list, and
path lists are separated by a semicolon on Windows. So before
passing this to git-peek-remote.exe, it transforms that to
"host;C:/msysGit/path". Avoid that by expanding it to
"ssh://host/path", but take extra care not to convert absolute
paths to that syntax!
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
This fix (while correct) actually avoids another nasty bug that must be fixed later:
environment.c caches results of many getenv calls.
Under MinGW setenv(X) invalidates all previous values returned by getenv(X)
so cached values become dangling pointers.
Signed-off-by: Dmitry Kakurin <Dmitry.Kakurin@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
This imitates what is currently in git.git's master: a function to
determine if a path is absolute.
msysgit.git is not really the proper place to fix this, but Hannes is
still on holiday.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
The expat build was added and the pieces of curl that seemed missing. The
default NO_CURL was removed so that curl is used by default.
Signed-off-by: Mike Pape <dotzenlabs@gmail.com>
It seems that you have to jump through hoops to make trivial things
such as "install.exe" do the right thing on Windows...
Also, it seems that Windows Vista deliberately broke "access()".
Apparently, they used that name in their runtime, but that function
does something completely different than POSIX access(). So define
__USE_MINGW_ACCESS to work around that.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
In setup_git_directory_gently(), there is a hack to allow for stopping
the search at the drive letter. Only that the patch was incomplete;
we really have to stop there, instead of looping infinitely.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
The MinGW port does have partial support for symbolic links now.
But in order to allow others who still do not have this version
to use their old version on Windows to pull and checkout this version
we better drop the symbolic link.
We do not yet support changing the directory of the spawned process
(it would require a getwd(), chdir(new), chdir(back) sequence).
Even though we can remove environment variables in the spawned process,
we cannot yet add new ones.
Move git-p4import.py and Documentation/git-p4import.txt into
a contrib/p4import directory. Add a README there directing
people to contrib/fast-import/git-p4 as a better alternative.
Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There were many operations that did not notice and report errors
to the CVS client, which would have resulted in corrupt working
tree.
Signed-off-by: Jim Meyering <jim@meyering.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* 'master' of git://repo.or.cz/git/fastimport:
Teach fast-import to recursively copy files/directories
Fix git-p4 on Windows to not use the Posix sysconf function.
Correct trivial typo in fast-import documentation
Make every builtin-*.c file #include "builtin.h".
Also takes care of some declaration/definition mismatches.
Signed-off-by: Peter Hagervall <hager@cs.umu.se>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some source material (e.g. Subversion dump files) perform directory
renames by telling us the directory was copied, then deleted in the
same revision. This makes it difficult for a frontend to convert
such data formats to a fast-import stream, as all the frontend has
on hand is "Copy a/ to b/; Delete a/" with no details about what
files are in a/, unless the frontend also kept track of all files.
The new 'C' subcommand within a commit allows the frontend to make a
recursive copy of one path to another path within the branch, without
needing to keep track of the individual file paths. The metadata
copy is performed in memory efficiently, but is implemented as a
copy-immediately operation, rather than copy-on-write.
With this new 'C' subcommand frontends could obviously implement an
'R' (rename) on their own as a combination of 'C' and 'D' (delete),
but since we have already offered up 'R' in the past and it is a
trivial thing to keep implemented I'm not going to deprecate it.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Add condition for Windows, since it doesn't support the os.sysconf module.
We hardcode the commandline limit to 2K, as that should work on most
Windows platforms.
Signed-off-by: Marius Storm-Olsen <marius@trolltech.com>
Acked-by: Simon Hausmann <simon@lst.de>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
We were previously sensitive to leading slashes in the fetch
lines and incorrectly writing them to the config if the user
used them (needlessly) in the command-line.
This fixes the issue and allows us to play nicely with legacy
configs that have leading slashes in fetch lines.
Thanks to Bradford Smith for figuring this out for me:
>
> This works:
>
> git-svn clone https://my.server.net/repos/path/ -Ttrunk/testing
> -ttags/testing -bbranches/testing testing
>
> This doesn't:
>
> git-svn clone https://my.server.net/repos/path -T/trunk/testing
> -t/tags/testing -b/branches/testing testing
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This script reads the existing commit log and .mailmap file,
and outputs author e-mail addresses that would map to more
than one names (most likely due to difference in the way they
are spelled, but some are due to ancient botched commits).
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace uses of cat that do nothing but writing the contents of
a single file to another command via pipe.
[jc: Original patch from Josh was somewhat buggy and rewrote
"cat $file | wc -l" to "wc -l $file", but this one should be Ok.]
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the previous behavior, "git-rm --cached" (without -f) had the same
restriction as "git-rm". This forced the user to use the -f flag in
situations which weren't actually dangerous, like:
$ git add foo # oops, I didn't want this
$ git rm --cached foo # back to initial situation
Previously, the index had to match the file *and* the HEAD. With
--cached, the index must now match the file *or* the HEAD. The behavior
without --cached is unchanged, but provides better error messages.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now, git-log family can take full range of internally supported date format
to their --date=<format> argument. Document it.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now we can use all internally supported date formats with
git log --date=<format>
syntax. Earlier, we only allowed relative/local/default.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These days, show_date() takes a date_mode parameter to specify
the output format, and a separate specialized function for dates
in E-mails does not make much sense anymore.
This retires show_rfc2822_date() function and make it just
another date output format.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Support output of full ISO 8601 style dates in e.g. git log
and other places that use interpolation for formatting.
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
