Subversion SVN Changelog

Move inclusion of svn_debug.h from svn_types.h to svn_private_config.h.

* subversion/include/svn_types.h: Don't include "private/svn_debug.h".

* configure.ac: Include "private/svn_debug.h" in generated svn_private_config.h.in
file.

* subversion/svn_private_config.hw: Include "private/svn_debug.h".

Review by: gstein

r40509: 2009-11-14 08:02:24 [lgo] Lieven Govaerts

/tools/buildbot/master/master.cfg [diff, log, file]

New configuration for the WanDisco CentOS buildslave.

* tools/buildbot/master/master.cfg: Add builder for x64-centos,
currently in test mode.

r40508: 2009-11-14 01:43:35 [julianfoad] Julian Foad

  [fs-successor-ids] /subversion/libsvn_fs_base/fs.c  [diff, log, file]
  [fs-successor-ids] /subversion/libsvn_fs_base/fs.h  [diff, log, file]
  [fs-successor-ids] /subversion/libsvn_fs_base/node-rev.c  [diff, log, file]
  [fs-successor-ids] /subversion/libsvn_fs_base/node-rev.h  [diff, log, file]
  [fs-successor-ids] /subversion/tests/cmdline/stat_tests.py  [diff, log, file]

On the fs-successor-ids branch: Bump the FS-BDB format number to 5, and make
the "successor ids" table conditional on format number at least 5.

* subversion/libsvn_fs_base/fs.c
(open_databases): Make opening the successor ids database conditional.

* subversion/libsvn_fs_base/fs.h
(SVN_FS_BASE__FORMAT_NUMBER): Bump from 4 to 5.
(SVN_FS_BASE__MIN_SUCCESSOR_IDS_FORMAT): New macro.

* subversion/libsvn_fs_base/node-rev.c
(svn_fs_base__create_successor, svn_fs_base__delete_node_revision): Make
the use of the successor ids table conditional.
(svn_fs_base__get_node_successors): Throw an "unsupported feature" error
if the format is less than 5.

* subversion/libsvn_fs_base/node-rev.h
(svn_fs_base__delete_node_revision): Fix a typo in the doc string.

r40507: 2009-11-13 20:18:38 [cmpilato] C. Michael Pilato

Click to show all 116 changes...

Sync up with the recent trunk changes. (Merged /trunk:r40393-40505)

r40506: 2009-11-13 20:16:23 [cmpilato] C. Michael Pilato

[issue-3242-dev] /subversion/tests/cmdline/authz_tests.py [diff, log, file]

Expect the unexpected. (Unexpected for good reason.)

* subversion/tests/cmdline/authz_tests.py
(test_list): Now expect authz_access_required_at_repo_root() to
pass. It only does over ra_serf, but, hey, one can hope.

r40505: 2009-11-13 19:56:02 [cmpilato] C. Michael Pilato

[issue-3242-dev] /subversion/libsvn_client/copy.c [diff, log, file]

Loosen up calculate_target_mergeinfo()'s restrictions, and fix its
callers.

* subversion/libsvn_client/copy.c
(calculate_target_mergeinfo): Drop the restriction about RA_SESSION's
parenting, and just reparent to the target for simplicity.
(wc_to_repos_copy, repos_to_wc_copy_single): Don't reparent the
session before calling calculate_target_mergeinfo(), and do pass
the right URL to calculate_target_mergeinfo().

r40504: 2009-11-13 19:36:17 [pburba] Paul T. Burba

/subversion/libsvn_client/merge.c [diff, log, file]

Improve error message for invalid revision keywords from a URL merge source.

* subversion/libsvn_client/merge.c
(normalize_merge_sources): Error out early when invalid revision keywords
are specified with a URL merge source rather than proceeding and
eventually returning a decidedly cryptic error -- see
http://svn.haxx.se/dev/archive-2009-11/0603.shtml.

Patch by: Kannan R <kannanr{_AT_}collab.net>
Tweaked by: me

r40503: 2009-11-13 18:24:30 [stsp] Stefan Sperling

/subversion/libsvn_client/patch.c [diff, log, file]

Using the diff3 algorithm in the patch code was a bad idea.
Diff3 isn't designed for this. It works nicely if all hunks apply cleanly,
but not in the general case where hunks might not match at all.
So remove use of svn_wc_merge4() in favour of a simple svn_io_copy_file()
of the the patched result into the working copy.

* subversion/libsvn_client/patch.c
(patch_target_t): Remove ORIGINAL, ORIGINAL_RAW, and ORIGINAL_PATH members.
We don't need the "original" stream anymore.
(init_patch_target): Don't open the original stream anymore.
(copy_lines_to_target): Don't write to the original stream anymore.
(apply_one_hunk): Don't copy hunk text to the original stream anymore.
Adjust a comment.
(apply_one_patch): Stop using svn_wc_merge4() to apply modifications to
the patch target. Just copy the patched temporary file on top of the target
instead. Adjust some comments.

r40502: 2009-11-13 17:44:22 [arfrever] Arfrever Frehtes Taifersar Arahesis

/build/generator/ezt.py [diff, log, file]

* build/generator/ezt.py: Support Python 3.

ezt upstream issue: http://code.google.com/p/ezt/issues/detail?id=5

r40501: 2009-11-13 17:42:31 [stsp] Stefan Sperling

/subversion/libsvn_client/patch.c [diff, log, file]

* subversion/libsvn_client/patch.c
(resolve_target_path): Remove the hack added in r40399 which was supposed
to help with applying patches containing paths with Windows-style path
separators on UNIX. The rationale is that UNIX patch does not do this,
so we shouldn't either. This commit does not revert r40399 entirely,
since we keep using svn_dirent_internal_style() rather than
svn_dirent_canonicalize() to canonicalize the target path parsed from
the patch file.

Suggested by: brane

r40500: 2009-11-13 17:38:43 [julianfoad] Julian Foad

  /tools/buildbot/slaves/centos  [log, file]
  /tools/buildbot/slaves/centos/svnbuild.sh  [log, file]
  /tools/buildbot/slaves/centos/svncheck-bindings.sh  [log, file]
  /tools/buildbot/slaves/centos/svncheck.sh  [log, file]
  /tools/buildbot/slaves/centos/svnclean.sh  [log, file]
  /tools/buildbot/slaves/centos/svnlog.sh  [log, file]

* tools/buildbot/slaves/centos
A new buildbot slave configuration.

* centos/svnbuild.sh
* centos/svncheck-bindings.sh
* centos/svncheck.sh
* centos/svnclean.sh
* centos/svnlog.sh
New scripts, modeled on the ubuntu ones.

r40499: 2009-11-13 15:38:31 [arfrever] Arfrever Frehtes Taifersar Arahesis

[python-3-compatibility] /build/run_tests.py [diff, log, file]

On the 'python-3-compatibility' branch:

* build/run_tests.py
(TestHarness.run): 'wimptag' should be str instead of bytes.

r40498: 2009-11-13 15:32:42 [arfrever] Arfrever Frehtes Taifersar Arahesis

/subversion/svn/cl.h [diff, log, file]

Follow-up to r40492:

* subversion/svn/cl.h
(svn_cl__path_join): Fix typo in doc string.

r40497: 2009-11-13 14:56:32 [pburba] Paul T. Burba

/subversion/libsvn_ra_neon/lock.c [diff, log, file]

Follow-up to r40493, fix a compiler warning.

* subversion/libsvn_ra_neon/lock.c
(do_lock): Remove unused format arg when creating an error.

r40496: 2009-11-13 14:36:32 [philip] Philip Martin

/Makefile.in [diff, log, file]

* Makefile.in: Remove misc-docs support as r40489 removed doc/misc-docs/.

r40495: 2009-11-13 11:57:51 [stsp] Stefan Sperling

/subversion/libsvn_client/status.c [diff, log, file]

Don't happily dump core upon svn status --ignore-externals, which we've
been doing since probably r39213.

* subversion/libsvn_client/status.c
(svn_client_status5): Flip mistaken boolean logic, we need to pass
an external_func + baton to svn_wc_walk_status() when *not* ignoring
externals.

r40494: 2009-11-13 09:02:52 [philip] Philip Martin

/subversion/libsvn_wc/upgrade.c [diff, log, file]

Allocate hashes before using them (gcc warns "may be used unitialized").

* subversion/libsvn_wc/upgrade.c (migrate_props): Allocate hashes.

r40493: 2009-11-13 02:37:57 [pburba] Paul T. Burba

  /subversion/libsvn_fs_base/lock.c  [diff, log, file]
  /subversion/libsvn_fs_fs/lock.c  [diff, log, file]
  /subversion/libsvn_ra_neon/lock.c  [diff, log, file]
  /subversion/libsvn_ra_serf/locks.c  [diff, log, file]
  /subversion/tests/cmdline/lock_tests.py  [diff, log, file]

Fix issue #3524 'Locking path via ra_serf which doesn't exist in HEAD
triggers assert'.

* subversion/libsvn_fs_base/lock.c (txn_body_lock):
* subversion/libsvn_fs_fs/lock.c (lock_body):
svn_fs_lock's API promises that if the "current_rev is a valid revnum,
then do an out-of-dateness check. If the revnum is less than the
last-changed-revision of path (or if path doesn't exist in HEAD), return
SVN_ERR_FS_OUT_OF_DATE." So let's actually start *doing* that for both
FS implementations.

* subversion/libsvn_ra_neon/lock.c (do_lock):
* subversion/libsvn_ra_serf/locks.c (handle_lock):
If a 405 error results from a LOCK request then return a
SVN_ERR_FS_OUT_OF_DATE error so ultimately the neon/serf svn_ra_lock
implementations will behave as promised by the API, i.e. "If the revnum
is less than the last-changed-revision of any path (or if a path doesn't
exist in HEAD), call lock_func/lock_baton with an
SVN_ERR_RA_OUT_OF_DATE error."

* subversion/tests/cmdline/lock_tests.py
(lock_path_not_in_head): Tweak test expectations to reflect that attempts
to lock paths that don't exist in HEAD are not fatal errors, see
SVN_ERR_IS_LOCK_ERROR.

r40492: 2009-11-12 23:37:51 [julianfoad] Julian Foad

  /subversion/svn/cl.h  [diff, log, file]
  /subversion/svn/diff-cmd.c  [diff, log, file]
  /subversion/svn/util.c  [diff, log, file]

Replace deprecated svn_path_join().

* subversion/svn/cl.h
(svn_cl__path_join): New.

* subversion/svn/util.c
(svn_cl__path_join): Wrap svn_path_join().

* subversion/svn/diff-cmd.c
(svn_cl__diff, summarize_xml, summarize_regular): Replace svn_path_join()
with svn_cl__path_join() since the resulting path is allowed to be both
uri and dirent.

Patch by: Daniel Näslund <daniel{_AT_}longitudo.com>
Suggested by: julianfoad

r40491: 2009-11-12 21:51:55 [stsp] Stefan Sperling

/subversion/svn/conflict-callbacks.c [diff, log, file]

* subversion/svn/conflict-callbacks.c
(svn_cl__conflict_handler): Add error message when trying to launch
an external merge tool for a property conflict.

Patch by: Daniel Näslund <daniel{_AT_}longitudo.com>
(error message tweaked by me)

Select page:

Previous Page 1 2 3 542 1082 1621

Subversion SVN Changelog

Select Changelog:

Select page:

Changes per page:

Summarize files:

Search:

Filter by Developer:

Changelog for Subversion (40,515 revisions):

Select page: