| Revision | Author | Date | Description |
|---|
| 35054 | cmpilato | 06 Jan 2009 15:22:23 | Fix some spelling oopses.
* subversion/tests/cmdline/svnsync_tests.py
s/syncron/synchron/
Found by: glasser
|
| 35053 | cmpilato | 06 Jan 2009 14:34:33 | Add an 'info' subcommand to svnsync to print the synchronization
information, and also add two tests for it.
* subversion/svnsync/main.c
(SVNSYNC_OPTS_DEFAULT): Remove 'q' from the default options.
(svnsync_cmd_table): Add the 'info' subcommand, and add 'q' to the
list of accepted options for the 'init', 'sync', and 'copy-revprops'
subcommands.
(info_cmd): New 'info' subcommand handler.
* subversion/tests/cmdline/svnsync_tests.py
(run_info): New helper function.
(info_syncronized, info_not_syncronized): New tests.
(test_list): Add references to new tests.
Patch by: Chris Foote
(Tweaked by me.)
|
| 35052 | julianfoad | 06 Jan 2009 13:58:34 | Fix an inefficiency in svn_stream_checksummed() that was introduced recently.
* subversion/libsvn_subr/stream.c
(svn_stream_checksummed): Remove redundant creation of checksum structures.
Don't ask svn_stream_checksummed2() to calculate checksums in all cases,
only when we want them. Add comments about what's going on.
|
| 35051 | julianfoad | 06 Jan 2009 13:46:07 | Simplify the signature of the new API svn_stream_checksummed2().
* subversion/include/svn_io.h,
subversion/libsvn_subr/stream.c
(svn_stream_checksummed2): Replace two separate checksum kind parameters
with just one, as it is unlikely that any user wants them to differ,
and if they do they can compose two checksumming streams.
* subversion/libsvn_fs/fs-loader.c
subversion/libsvn_subr/io.c
subversion/libsvn_subr/stream.c
subversion/libsvn_wc/adm_crawler.c
subversion/libsvn_wc/questions.c
subversion/libsvn_wc/update_editor.c
subversion/tests/libsvn_fs/fs-test.c
Adjust all calls to it.
|
| 35050 | julianfoad | 06 Jan 2009 13:15:33 | Tweak a local function signature to avoid the need for a type cast. (This
particular type cast was casting away "const" though it need not have done.)
* subversion/libsvn_subr/svn_base64.c
(encode_bytes): Take "void *" instead of "char *" input data.
(svn_base64_from_checksum): Remove a type cast in calling encode_bytes().
|
| 35049 | pburba | 06 Jan 2009 12:11:11 | Follow-up to r34863, stop spurious passing of all tests on Windows.
Found by: rhuijben
See http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&
dsMessageId=993668.
* subversion/tests/cmdline/svntest/main.py
(open_pipe): Allow subprocess.Popen() to successfully run a python script
on Windows by ensuring the first arg is the Python executable, not the
script name.
|
| 35047 | fabien | 06 Jan 2009 11:33:43 | French translation update for r35045
* subversion/po/fr.po: po-update, one new and a few fuzzy fixed
|
| 35046 | stsp | 06 Jan 2009 11:29:31 | * TODO-1.6: Add #3354 as blocker.
|
| 35045 | cmpilato | 06 Jan 2009 11:13:20 | * subversion/svnadmin/main.c
(pack_notify): Build a formatted string separately so gettext can do
the right thing for the final target string.
Found by: Fabien COELHO |
| 35044 | cmpilato | 06 Jan 2009 11:11:05 | Minor optimization to mergeinfo parsing.
* subversion/libsvn_subr/mergeinfo.c
(parse_pathname): Avoid unnecessary reallocation by simply creating
a stringbuf fully in a single step.
|
| 35040 | pburba | 06 Jan 2009 10:31:29 | Allow colons within mergeinfo path names.
Previously a mergeinfo pathname containing colons produced a parsing error.
Found by: Damien Deville
* subversion/libsvn_subr/mergeinfo.c
(parse_pathname):
* subversion/tests/libsvn_subr/mergeinfo-test.c
(NBR_MERGEINFO_VALS, mergeinfo_vals, mergeinfo_paths, mergeinfo_ranges):
Add some test cases for test_parse_single_line_mergeinfo() which have
pathnames with colons.
(NBR_BROKEN_MERGEINFO_VALS, broken_mergeinfo_vals):
Add some test cases for test_parse_broken_mergeinfo() which have
pathnames with colons.
(test_merge_mergeinfo): Add some colons to pathnames. |
| 35039 | cmpilato | 06 Jan 2009 10:19:55 | * tests/libsvn_fs_base/fs-base-test.c
(orphaned_textmod_change): Add a comment explaining why we have this
test at all, given that it violates the API rules.
Suggested by: glasser |
| 35037 | gstein | 06 Jan 2009 06:07:20 | axed! |
