Package: subversion / 1.10.4-1+deb10u3
Metadata
Package | Version | Patches format |
---|---|---|
subversion | 1.10.4-1+deb10u3 | 3.0 (quilt) |
Patch series
view the series filePatch | File delta | Description |
---|---|---|
build fixes | (download) |
Makefile.in |
8 3 + 5 - 0 ! |
build-fixes Fix some build problems: Fix some build problems: - Add aprutil as a direct dependency for libsvn_client and libsvn_ra. |
last changed date charset | (download) |
subversion/libsvn_subr/subst.c |
8 6 + 2 - 0 ! |
last-changed-date-charset Bug #290774: Convert $LastChangedDate$ keyword to the local character Bug #290774: Convert $LastChangedDate$ keyword to the local character set, since it is already rendered in the local language. Not accepted upstream: http://svn.haxx.se/dev/archive-2006-04/0730.shtml |
repos templates | (download) |
subversion/libsvn_repos/repos.c |
28 18 + 10 - 0 ! |
repos-templates Bug #210901: Add fully qualified paths to commands in example hook Bug #210901: Add fully qualified paths to commands in example hook scripts that get installed as .tmpl files in new repositories. Use $REPOS/hooks/ as the path, to encourage repository admins to copy the hooks, not just use them in place. |
rpath | (download) |
build.conf |
4 2 + 2 - 0 ! |
rpath Prevent the linker from adding an rpath to shared libraries. Original patch by David Kimdon <dwhedon@debian.org>. The basic theory is: - Use libtool instead of apxs to install the apache modules. libtool relinks without rpath in this case, apxs obviously doesn't. |
ruby no strict aliasing | (download) |
Makefile.in |
2 1 + 1 - 0 ! |
ruby-no-strict-aliasing It seems swig is generating _really_ bad code here, or something. Whatever it is, this line suppresses 1323 warnings, any of which might be real bugs with real consequences. |
ruby test svnserve race | (download) |
subversion/bindings/swig/ruby/test/util.rb |
28 23 + 5 - 0 ! |
ruby-test-svnserve-race Bug #378837: Ruby testsuite: wait for svnserve to start before Bug #378837: Ruby testsuite: wait for svnserve to start before connecting to it. This mainly affects very slow machines - observed on various arm and m68k builds. Thanks to Roman Zippel, Kobayashi Noritada, Wouter Verhelst and Martin Michlmayr. |
ssh no controlmaster | (download) |
subversion/libsvn_ra_svn/client.c |
2 1 + 1 - 0 ! |
ssh-no-controlmaster Bug #413102: Avoid accidentally becoming an ssh ControlMaster. Bug #413102: Avoid accidentally becoming an ssh ControlMaster. If you set ControlMaster via your ssh config: .ssh/config: Host * ControlMaster auto Controlpath ~/.ssh/master-%r@%h:%p you can get 'svn: Can't write to connection: Broken pipe' with certain ssh commands, because svn can open multiple ssh sessions and kill them at unpredictable times. I don't really think the problem is either svn's or openssh's fault, but this is the obvious way to avoid it. We will still use ssh connection sharing, but only if a ControlMaster is already present. |
svnshell | (download) |
tools/examples/svnshell.py |
8 1 + 7 - 0 ! |
svnshell Bug #234462: The error messages that svnshell spits out for unknown Bug #234462: The error messages that svnshell spits out for unknown commands are confusing. They should be replaced by a "Unknown command." error message. Upstream says we have no sense of humor. (: Patch from Roland Illig <roland.illig@gmx.de>. |
apr abi | (download) |
configure.ac |
2 1 + 1 - 0 ! |
apr-abi Bug #266170: Use SOVERSION 1 for apr 1.x, because the apr 0.9.x -> 1.x Bug #266170: Use SOVERSION 1 for apr 1.x, because the apr 0.9.x -> 1.x ABI change leaks through the libsvn ABIs (because we use apr_off_t, which is now 64 bits on all platforms). Upstream refused this (http://svn.haxx.se/dev/archive-2006-03/0522.shtml for thread), but fortunately when it came time to support apr 2.x in svn 1.7, they skipped SOVERSION 1 in consideration of Debian/Ubuntu, so we'll be ABI-compatible again if and when we move to apr-2.x. |
python builds | (download) |
Makefile.in |
1 1 + 0 - 0 ! |
python-builds Build for multiple python versions. Patch from Matthias Klose <doko@debian.org> via Ubuntu. Mostly, rename libsvn_swig_py to libsvn_swig_py2.x. Submitted upstream: http://svn.haxx.se/dev/archive-2008-05/0361.shtml |
examples compile instructions | (download) |
tools/examples/getlocks_test.c |
8 3 + 5 - 0 ! |
examples-compile-instructions Fix build instructions for libsvn-doc/examples/*.c. |
workaround_EINVAL_on_kfreebsd | (download) |
subversion/bindings/swig/ruby/test/util.rb |
2 2 + 0 - 0 ! |
workaround_einval_on_kfreebsd |
java10 compatibility | (download) |
build/ac-macros/java.m4 |
2 1 + 1 - 0 ! |
java10-compatibility |
0014 Provide a way for svnserve s get deleted rev API to .patch | (download) |
subversion/libsvn_ra_svn/client.c |
17 15 + 2 - 0 ! |
provide a way for svnserve's 'get-deleted-rev' api to return 'not deleted'. Previously the answer 'not deleted' was indistinguishable from an error when using the svnserve protocol. The standard 'svn' client software does not appear to invoke this case, but it is meant to be a valid answer and is supported by the other repository access protocols. To work around the problem without changing the protocol, we re-purpose the obsolete error code 'SVN_ERR_ENTRY_MISSING_REVISION' to communicate this 'not deleted' reply to the client. (Re-purposing an old error code was not essential. An alternative would be to create a new error code which would become part of the public API in the next minor release, but would be deemed non-public when backported to older release lines.) Behaviour changes: - With a new client against a new server, such queries are now handled correctly. - With an old client against a new server, the client will report a more informative error message. - With a new client against an old server, there is no improvement. * subversion/svnserve/serve.c (get_deleted_rev): If the answer is SVN_INVALID_REVNUM, return a SVN_ERR_ENTRY_MISSING_REVISION error to the client. * subversion/libsvn_ra_svn/client.c (ra_svn_get_deleted_rev): Convert the error SVN_ERR_ENTRY_MISSING_REVISION back to a response of SVN_INVALID_REVNUM. * subversion/tests/libsvn_ra/ra-test.c (commit_two_changes): New. (test_get_deleted_rev_no_delete, test_get_deleted_rev_errors): New tests. (test_funcs): Run them. Signed-off-by: James McCoy <jamessan@debian.org> |
0015 Adjust expectations of a new test to account for dif.patch | (download) |
subversion/tests/libsvn_ra/ra-test.c |
13 10 + 3 - 0 ! |
adjust expectations of a new test to account for differences between RA layers. A follow-up to r1860936. * subversion/tests/libsvn_ra/ra-test.c (test_get_deleted_rev_errors): Expect a generic error code when running over HTTP, otherwise the specific error code for this error case. Signed-off-by: James McCoy <jamessan@debian.org> |
0016 In svnserve consistently handle errors in opening a .patch | (download) |
subversion/svnserve/serve.c |
13 7 + 6 - 0 ! |
in svnserve, consistently handle errors in opening a repository. These errors are still logged and reported to the client, as they were before, but now it is done in the same way as everywhere else. (The error logging now happens higher up the call stack.) * subversion/svnserve/serve.c After reporting an error to the client, don't log it explicity here and then clear it; instead return it so the caller can do so. Signed-off-by: James McCoy <jamessan@debian.org> |
0017 Fix a potential NULL dereference in the config file .patch | (download) |
subversion/libsvn_repos/config_file.c |
4 4 + 0 - 0 ! |
fix a potential null dereference in the config file parser. * subversion/libsvn_repos/config_file.c (get_repos_config): svn_repos_find_root_path() may return NULL. Check the return value accordingly. |
0018 Don t show unreadable copyfrom paths in svn log v.patch | (download) |
subversion/libsvn_repos/log.c |
26 10 + 16 - 0 ! |
don't show unreadable copyfrom paths in 'svn log -v' * subversion/libsvn_repos/log.c: (detect_changed): In the add and replace cases, explicitly clear the svn_fs_path_change3_t::copyfrom_path pointer and rev when said path@rev is unreadable. * subversion/tests/cmdline/authz_tests.py: (log_inaccessible_copyfrom): New regression test. (test_list): Run it. Patch by: kotkov git-svn-id: https://svn.apache.org/repos/asf/subversion/trunk@1899227 13f79535-47bb-0310-9956-ffa450edef68 Signed-off-by: James McCoy <jamessan@debian.org> |
0019 Fix issue 4880 Use after free of object pools when u.patch | (download) |
subversion/libsvn_repos/authz.c |
25 25 + 0 - 0 ! |
fix issue #4880 "use-after-free of object-pools when used as httpd module" MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Ensure that we initialize authz again if the pool which our authz caches depend on is cleared. Apache HTTPD may run pre/post config hooks multiple times and clear its global configuration pool which our authz caching pools depend on. Reported-by: Thomas Weischuh (thomas {at} t-8ch dot de) Thomas has also confirmed that this patch fixes the problem. * subversion/libsvn_repos/authz.c (deinit_authz): New pool cleanup handler which resets authz initialization in case the parent pool of our authz caches is cleared. (synchronized_authz_initialize): Register new pool cleanup handler. git-svn-id: https://svn.apache.org/repos/asf/subversion/trunk@1894734 13f79535-47bb-0310-9956-ffa450edef68 Signed-off-by: James McCoy <jamessan@debian.org> |