Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 23 Mar 2006 23:55:51 +0100] rev 1991
merge with self
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 23 Mar 2006 23:39:53 +0100] rev 1990
make it possible to escape characters in a glob expression
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 23 Mar 2006 23:16:41 +0100] rev 1989
ignore EPIPE in ui.err_write
It avoids not being able to abort a transaction when a push via ssh fails.
Maybe some other place should ignore EPIPE too.
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 21 Mar 2006 23:31:04 -0800] rev 1988
merge with crew.
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 21 Mar 2006 23:29:21 -0800] rev 1987
add changelog style to command line template.
to use, "hg log --style=changelog". makes different output with no
flags, -q, -v, --debug.
templater module has new template filters for this.
email - committer email address
fill68 - refill text to 68 colums
fill76 - refill text to 76 colums
tabindent - prefix every not empty line with tab
shortdate - iso 8631 date, no time zone
stringify - turn template iterator into string
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Wed, 22 Mar 2006 05:30:47 +0100] rev 1986
add checking for invalid entries in tag files
safely parse the differents tag files, output warning
when parsing an invalid entry.
Thomas Arendsen Hein <thomas@intevation.de> [Tue, 21 Mar 2006 15:33:29 +0100] rev 1985
Adapted behaviour of ui.username() to documentation and mention it explicitly:
Searched in this order: $HGUSER, [ui] section of hgrcs, $EMAIL
and stop searching if one of these is set.
Abort if found username is an empty string to force specifying
the commit user elsewhere, e.g. with line option or repo hgrc.
If not found, use $LOGNAME or $USERNAME +"@full.hostname".
Thomas Arendsen Hein <thomas@intevation.de> [Tue, 21 Mar 2006 12:54:32 +0100] rev 1984
Improved ui.edit():
- Use descriptive name for temporary file: hg-editor-*.txt
- Don't import tempfile in the method, but use demandload()
- Remove tempfile file even if editor aborts.
Thomas Arendsen Hein <thomas@intevation.de> [Tue, 21 Mar 2006 12:45:27 +0100] rev 1983
Pass correct username as $HGUSER to hgeditor if "commit -u" is used.
And only use the new util.system parameter to set the environment.
Thomas Arendsen Hein <thomas@intevation.de> [Tue, 21 Mar 2006 12:15:25 +0100] rev 1982
Added test for incoming via ssh.
Thomas Arendsen Hein <thomas@intevation.de> [Tue, 21 Mar 2006 11:47:21 +0100] rev 1981
make incoming work via ssh (issue139); move chunk code into separate module.
Incoming ssh needs to detect the end of the changegroup, otherwise it would
block trying to read from the ssh pipe. This is done by parsing the
changegroup chunks.
bundlerepo.getchunk() already is identical to
localrepo.addchangegroup.getchunk(), which is followed by getgroup which
looks much like what you can re-use in bundlerepository.__init__() and in
write_bundle(). bundlerevlog.__init__.genchunk() looks very similar, too,
as do some while loops in localrepo.py.
Applied patch from Benoit Boissinot to move duplicate/related code
to mercurial/changegroup.py and use this to fix incoming ssh.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Tue, 21 Mar 2006 06:03:33 +0100] rev 1980
use HG10UN header for uncompressed bundle
- use HG10UN instead of HG11 for uncompressed bundles header
- check HG10BZ for compressed bundle
- better error handling for invalid header
some notes:
- people who created uncompressed bundle will no longer be
able to use them (it could be fixed with hand-editing)
- older hg cannot detect an uncompressed bundle
(bzip2 decompression will fail).
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 20 Mar 2006 17:14:58 +0100] rev 1979
More detailed documentation about ssh:// URLs; fixes issue170.
TK Soh <teekaysoh@yahoo.com> [Mon, 20 Mar 2006 08:46:29 +0100] rev 1978
do proper typecasting on malloc() and calloc() calls
to support build on Solaris 2.6 using Sun Pro SC4.0 (C++ 4.1) compiler.
Thomas Arendsen Hein <thomas@intevation.de> [Sun, 19 Mar 2006 21:26:58 +0100] rev 1977
Don't forget version at the end of setup.py, write it only if changed.
This fixes issue159: "python setup.py install" shouldn't write new files
in the working dir if one has done "python setup.py build" first.
Thomas Arendsen Hein <thomas@intevation.de> [Sun, 19 Mar 2006 14:53:58 +0100] rev 1976
Enable path validation for copy, rename, debugwalk and other canonpath users.
This fixed the last bit of issue134.
Added some tests to test-rename and test-walk.
Shun-ichi Goto <shunichi.goto@gmail.com> [Sat, 18 Mar 2006 09:51:25 -0800] rev 1975
Duplicate cache when creating templater.
Thomas Arendsen Hein <thomas@intevation.de> [Sat, 18 Mar 2006 14:36:45 +0100] rev 1974
Delete bundle file of hg incoming in case of errors, preserve existing files.
Let write_bundle always create the bundle file, check if target doesn't exist
and delete it in case of errors.
This way incoming only has to delete the bundle file if it was meant to
be a temporary file.
Thomas Arendsen Hein <thomas@intevation.de> [Sat, 18 Mar 2006 11:24:19 +0100] rev 1973
Cleaned up the code path for bundle handling in hg incoming, updated comments.
Thomas Arendsen Hein <thomas@intevation.de> [Sat, 18 Mar 2006 10:45:21 +0100] rev 1972
Documented behaviour of write_bundle regarding file handles.
Thomas Arendsen Hein <thomas@intevation.de> [Sat, 18 Mar 2006 09:13:27 +0100] rev 1971
imported patch /home/thomas/fix-incoming-abortion4.patch
Shun-ichi GOTO <shunichi.goto@gmail.com> [Fri, 17 Mar 2006 17:23:52 -0800] rev 1970
Fixed abort while remote incoming command on windows.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Fri, 17 Mar 2006 18:53:04 +0100] rev 1969
write_bundle: close the file before removing it in case of failure
Vadim Gelfer <vadim.gelfer@gmail.com> [Fri, 17 Mar 2006 09:49:28 -0800] rev 1968
merge with crew.
Vadim Gelfer <vadim.gelfer@gmail.com> [Fri, 17 Mar 2006 09:48:43 -0800] rev 1967
still use old variable names when running hooks. dropped by accident.
we set HG_FOO (new name) and FOO (old name).
plan is to drop old names after 0.9.
Thomas Arendsen Hein <thomas@intevation.de> [Fri, 17 Mar 2006 18:19:05 +0100] rev 1966
hg status cleanups: Don't translate long options, remove stray semicolon.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 16 Mar 2006 06:58:15 +0100] rev 1965
add remotecmd/ssh option support to incoming/outgoing
Vadim Gelfer <vadim.gelfer@gmail.com> [Wed, 15 Mar 2006 11:28:00 -0800] rev 1964
fix template bug that made hgweb break.
problem found by Shun-ichi Goto <shunichi.goto@gmail.com>.
this fix tested with command template and hgweb templates.
Thomas Arendsen Hein <thomas@intevation.de> [Wed, 15 Mar 2006 11:26:57 +0100] rev 1963
Don't require the "yes" command for test-merge5.
TK Soh <teekaysoh@yahoo.com> [Wed, 15 Mar 2006 11:09:24 +0100] rev 1962
tests: more portability fix for the old systems
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 14 Mar 2006 23:01:11 -0800] rev 1961
merge with crew.
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 14 Mar 2006 23:00:45 -0800] rev 1960
update test for unrelated repo.
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 14 Mar 2006 22:58:14 -0800] rev 1959
add -f/--force to pull, incoming, outgoing, to work on unrelated repo.
before this, push would not push from e.g. "hg" repo to "kernel" repo
but other commands worked. this was bad idea, could merge unrelated
projects by accident. i did this tonight.
now, all commands still work with unrelated repo but need
--force/-f. abort is default. this is safer.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Wed, 15 Mar 2006 07:15:13 +0100] rev 1958
merge with crew
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Wed, 15 Mar 2006 07:10:04 +0100] rev 1957
merge with crew
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Wed, 15 Mar 2006 07:09:42 +0100] rev 1956
use a proper test instead of catching every exception
Shun-ichi GOTO <shunichi.goto@gmail.com> [Wed, 15 Mar 2006 06:39:04 +0100] rev 1955
Fix the py2exe template path problem (issue:152) again.
This patch is same with 1834:24881eaebee3 which is leaved behind by
1897:58b6784cf9f1.
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 14 Mar 2006 22:02:41 -0800] rev 1954
merge with crew.
TK Soh <teekaysoh@yahoo.com> [Tue, 14 Mar 2006 22:01:30 -0800] rev 1953
change default strftime format for better portability
On some systems, such as Solaris and HP-UX, "%c" format
give output without the leading zero on the single-digit
month-day.
TK Soh <teekaysoh@yahoo.com> [Tue, 14 Mar 2006 21:42:32 -0800] rev 1952
tests: fix portability issue for echo -n
Vadim Gelfer <vadim.gelfer@gmail.com> [Tue, 14 Mar 2006 21:40:46 -0800] rev 1951
add HGRCPATH env var, list of places to look for hgrc files.
if set, override default hgrc search path.
if empty, only .hg/hgrc of current repo read.
for each element, if directory, all entries in directory with end in
".rc" are added to path. else, element is added to path.
big thing about this change is that user "~/.hgrc" and system hgrc not
longer breaks tests. run-tests makes HGRCPATH empty now.
Hollis Blanchard <hollisb@us.ibm.com> [Wed, 15 Mar 2006 03:19:16 +0100] rev 1950
hgk "committer:" bug
I've been having an hgk problem with this tree:
http://xenbits.xensource.com/ext/linux-ppc-2.6.hg, specifically changeset
93c590d23a53.
The problem seems to be that the commit message contains a "committer:" line,
which triggers a (relatively unused?) case in hgk.py... Both cases need the
dates at the end of the line.
Soh Tk-r28629 <tksoh@freescale.com> [Wed, 15 Mar 2006 02:49:45 +0100] rev 1949
tests: add bunzip2 to prerequisite list
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 23:38:14 +0100] rev 1948
merge with crew/templates
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 23:32:57 +0100] rev 1947
fix the call to util.matcher (the args should be (reporoot, cwd, ...))
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 06:58:31 +0100] rev 1946
bundlerepo: remove relative import, fix a comment
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:56:36 +0100] rev 1945
new type of repo: bundle://path/to/repo+/path/to/bundlename
You can use it to pull only some changeset from an uncompressed
bundle.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:54:23 +0100] rev 1944
incoming: add support for remote repo using bundlerepo
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:54:23 +0100] rev 1943
refactor the bundle writing code, since we will reuse it later
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:54:23 +0100] rev 1942
add bundlerepo.py: a read-only repo that can use uncompressed bundles
The idea is to create a repo consisting of a normal local repository
plus all the changesets contained in a bundle.
The bundle needs to be uncompressed. A futur version could implement
the seeking through a compressed bundle.
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:54:23 +0100] rev 1941
revlog.py: factorization and fixes for rev < 0 (nullid)
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Mon, 13 Mar 2006 03:54:23 +0100] rev 1940
add a new bundle type: uncompressed bundle
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 18:33:33 +0100] rev 1939
Use timestamps without leading 0 in test-command-template for portable tests.
Vadim Gelfer <vadim.gelfer@gmail.com> [Mon, 13 Mar 2006 08:25:33 -0800] rev 1938
merge with crew.
Vadim Gelfer <vadim.gelfer@gmail.com> [Mon, 13 Mar 2006 08:24:44 -0800] rev 1937
fix use of python 2.4 features.
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 13:19:23 +0100] rev 1936
Make test-remove portable for strange strftime implementations (second try).
I forgot to replace the date output for 'hg log -p', too.
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 13:11:35 +0100] rev 1935
Removed obsolete workaround for error message in tests/test-static-http.
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 13:08:31 +0100] rev 1934
Make test-remove portable for strange strftime implementations.
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 13:05:41 +0100] rev 1933
Use 'hg ci -d "1000000 0"' in tests to circumvent problem with leading zero.
Some systems show "Thu Jan 01" instead of "Thu Jan 1", which breaks tests.
Using "1000000" yields "Mon Jan 12 13:46:40 1970", which looks the same on
all systems.
Thomas Arendsen Hein <thomas@intevation.de> [Mon, 13 Mar 2006 12:22:55 +0100] rev 1932
Merge with http://hg.home.dataloss.nl/hg-portabletests