The standard allocator on Windows is pretty bad prior
to Windows Vista, and nedmalloc is better than the
modified dlmalloc provided with newer versions of the
MinGW libc.
NedMalloc stats in Git
----------------------
All results are the best result out of 3 runs. The
benchmarks have been done on different hardware, so
the repack times are not comparable.
These benchmarks are all based on 'git repack -adf'
on the Linux kernel.
XP
-----------------------------------------------
MinGW Threads Total Time Speed
-----------------------------------------------
3.4.2 (1T) 00:12:28.422
3.4.2 + nedmalloc (1T) 00:07:25.437 1.68x
3.4.5 (1T) 00:12:20.718
3.4.5 + nedmalloc (1T) 00:07:24.809 1.67x
4.3.3-tdm (1T) 00:12:01.843
4.3.3-tdm + nedmalloc (1T) 00:07:16.468 1.65x
4.3.3-tdm (2T) 00:07:35.062
4.3.3-tdm + nedmalloc (2T) 00:04:57.874 1.54x
Vista
-----------------------------------------------
MinGW Threads Total Time Speed
-----------------------------------------------
4.3.3-tdm (1T) 00:07:40.844
4.3.3-tdm + nedmalloc (1T) 00:07:17.548 1.05x
4.3.3-tdm (2T) 00:05:33.746
4.3.3-tdm + nedmalloc (2T) 00:05:27.334 1.02x
Mac Mini
-----------------------------------------------
GCC Threads Total Time Speed
-----------------------------------------------
i686-darwin9-4.0.1 (2T) 00:09:57.346
i686-darwin9-4.0.1+ned (2T) 00:08:51.072 1.12x
Signed-off-by: Marius Storm-Olsen <marius@trolltech.com>
The function FillConsoleOutputCharacterA() was pretty content in XP to take a NULL
pointer if we did not want to store the number of written columns. In Windows 7,
it crashes, but only when called from within Git Bash, not from within cmd.exe.
Go figure.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
As recorded in msysGit issue 125 if the user closes gitk while it
reports itself as still reading references then Tk will crash in the
geometry management code. This has been fixed for Tk 8.5.7 and above.
This patch avoids the problem by flushing outstanding geometry events
before calling the readrefs procedure.
See also http://code.google.com/p/msysgit/issues/detail?id=125
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
This reverts the following commits:
abd1fa1 test-suite: Log everything to a file in non-verbose mode
50f37d1 Implement test-stat the prints the modification time.
fc7fe85 gitk: Swap positions of 'next' and 'prev' buttons in the 'Find' section.
f80f1db Need diff -u -b in t7401 because some lines end in CRLF.
85763c8 Implement thread-specific die() routines; use one in start_async().
4eb0463 Make report() from usage.c public as vreport().
ddce705 Skip tests that fail due to incomplete implementations, missing tools...
a094080 Do not issue the warning about the fallback of the PREFIX.
These were brought in by the merge commit dc8b641 that merged my private
work-in-progress branch. Only these two commits remain:
1e52e22 Windows: Work around intermittent failures in mingw_rename
ccd3859 Windows: Better support PAGER settings with spaces in the path
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
TortoiseMerge comes with TortoiseSVN or TortoiseGit for Windows. It can
only be used as a merge tool with an existing base file. It cannot be
used without a base nor as a diff tool.
The documentation only mentions the slash '/' as command line option
prefix, which refused to work, but the parser also accepts the dash '-'
See http://code.google.com/p/msysgit/issues/detail?id=226
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Under Windows vimdiff and gvimdiff are not available as symbolic links,
but as batch files vimdiff.bat and gvimdiff.bat. These files weren't
found by 'type vimdiff' which led to the following error:
The merge tool vimdiff is not available as 'vimdiff'
Even if they were found, it wouldn't work to invoke these batch files
from git-mergetool.
To solve this, use vim and gvim (vim.exe and gvim.exe) and pass the -d
command line switch over to them.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
We have replaced rename() with a version that can rename a file to a
destination that already exists. Nevertheless, many users, the author
included, observe failures in the code that are not reproducible.
The theory is that the failures are due to some other process that happens
to have opened the destination file briefly at the wrong moment. (And there
is no way on Windows to delete or replace a file that is currently open.)
One such process could be Windows Explorer that is updating its directory
listing because it detected a change in the .git directory or a virus
scanner. The failure is more often observed while there is heavy git
activity (for example while the test suite is running or during a rebase
operation).
We work around the failure by retrying the rename operation if it failed
due to ERROR_ACCESS_DENIED. The retries are delayed a bit: The first only
by giving up the time slice, the next after the minimal scheduling
granularity, and if more retries are needed, then we wait some non-trivial
amount of time with exponential back-off.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Using 'return' from a test snippet is now a no-go because it would skip the
assignment of the resulting exit code and it would also skip set +x.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
The button order 'prev' left of 'next' feels more natural than the other
way round, in particular, compared to the order of the forward and backward
arrows that are in the line above.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at>
On Windows, the pager was always run via a shell, which is not the case
on Unix, where the shell variant was only used as a fallback.
Thus, setting the pager, for example, like this:
PAGER="C:/Program Files/msys/bin/less"
would fail; an extra set of quotes is needed:
PAGER="\"C:/Program Files/msys/bin/less\""
With this patch the former setting works as well.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at>
* dm/maint-docco:
Documentation: Remove spurious uses of "you" in git-bisect.txt.
Documentation: minor grammatical fix in git-check-ref-format.txt
Documentation: minor grammatical fixes in git-check-attr.txt
Documentation: minor grammatical fixes in git-cat-file.txt
Documentation: minor grammatical fixes and rewording in git-bundle.txt
Documentation: remove some uses of the passive voice in git-bisect.txt
Documentation: reword example text in git-bisect.txt.
Documentation: reworded the "Description" section of git-bisect.txt.
Documentation: minor grammatical fixes in git-branch.txt.
Documentation: minor grammatical fixes in git-blame.txt.
Documentation: reword the "Description" section of git-bisect.txt.
Documentation: minor grammatical fixes in git-archive.txt.
sanitize_address assumes that quoted addresses (e.g., "first last"
<first.last@example.com) do not need rfc2047 encoding, but this is
not always the case.
For example, various places in send-email extract addresses using
parse_address_line. parse_address_line returns the addresses already
quoted (e.g., "first last" <first.last@example.com), but not rfc2047
encoded.
This patch makes sanitize_address stricter about what needs rfc2047
encoding and adds a test demonstrating where I noticed the problem.
Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit c18f75a (send-email: add tests for refactored prompting, 2009-03-28)
added two tests which went interactive under the dash shell.
This patch corrects the issue, reported by Björn Steinbrink.
Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* ef/fast-export:
builtin-fast-export.c: handle nested tags
builtin-fast-export.c: fix crash on tagged trees
builtin-fast-export.c: turn error into warning
test-suite: adding a test for fast-export with tag variants
* cj/doc-format:
Documentation: use "spurious .sp" XSLT if DOCBOOK_SUPPRESS_SP is set
Documentation: option to render literal text as bold for manpages
Documentation: asciidoc.conf: fix verse block with block titles
Documentation: asciidoc.conf: always use <literallayout> for [blocktext]
Documentation: move "spurious .sp" code into manpage-base.xsl
Documentation: move quieting params into manpage-base.xsl
Documentation: rename docbook-xsl-172 attribute to git-asciidoc-no-roff
Documentation: use parametrized manpage-base.xsl with manpage-{1.72,normal}.xsl
Documentation: move callouts.xsl to manpage-{base,normal}.xsl
Documentation/Makefile: break up texi pipeline
Documentation/Makefile: make most operations "quiet"
* jc/maint-1.6.0-keep-pack:
pack-objects: don't loosen objects available in alternate or kept packs
t7700: demonstrate repack flaw which may loosen objects unnecessarily
Remove --kept-pack-only option and associated infrastructure
pack-objects: only repack or loosen objects residing in "local" packs
git-repack.sh: don't use --kept-pack-only option to pack-objects
t7700-repack: add two new tests demonstrating repacking flaws
Conflicts:
t/t7700-repack.sh
Fix the git-svn documentation svn-remote example section talking about
tags and branches by using the proper key "fetch" instead of "trunk".
Using "trunk" actually might be nice, but it doesn't currently work.
The fetch line for the trunk was also reordered to be at the top of the
list, since most people think about the trunk/tags/branches trio in that
logical order.
Signed-off-by: Wesley J. Landaker <wjl@icecavern.net>
Acked-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While it makes no sense to map some email address to an empty one, doing
things the other way around can be useful. For example when using
filter-branch with an env-filter that employs a mailmap to fix up an
import that created such broken commits with empty email addresses.
Signed-off-by: Björn Steinbrink <B.Steinbrink@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 6e18251 made the "Send this email?" prompt assume yes if confirm
= "inform" when it was unable to get a valid response. However, the
"yes" assumption only worked correctly for the first email. This commit
fixes the issue and confirms the fix by modifying the existing test for
the prompt to send multiple emails.
Reported by Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Attempting to prompt when the tty is closed (typically when running from
cron) is pointless and emits a warning. This patch causes ask() to
return early, squelching the warning.
Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some old versions of gcc don't seem to like us negating an
enum constant. Let's work around it by negating the other
half of the comparison instead.
Reported by Pierre Poissinger on gcc 2.9.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
