* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * THIS RELEASE STREAM IS OPEN TO BUG FIXES. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * This file tracks the status of releases in the 1.7.x line. See http://subversion.apache.org/docs/community-guide/releasing.html#release-stabilization for details on how release lines and voting work, what kinds of bugs can delay a release, etc. Status of 1.7.9: Candidate changes: ================== * r1306275, and other revisions New diff support code for added/deleted files and directories. Fixes issue #4153, "svn log --diff" on moved file gives "not found", and other problems. Justification: The current code (introduced in 1.7.x in the r1207351 backport commit) produces inconsistent paths in diff output (see http://svn.haxx.se/dev/archive-2012-03/0385.shtml), and it also breaks with the paradigm that the target of an editor drive may never have more than one path component (which according to cmpilato is an API violation). Notes: Please read the branch and related trunk log messages for detailed notes about these changes. Branch: ^/subversion/branches/1.7.x-issue4153 Votes: +1: stsp * ^/subversion/branches/1.7.x-commit-performance r1337118, r1337281, r1337374, r1337377, r1337392, r1337394, r1337485 Use the working copy status walker as a first filter to find committable nodes. Justification: Huge performance boost. Especially on NFS. Notes: Backports the status walker framework. Many other parts are backportable too, but this branch delivers the plumbing. [JAF] We might want to include some of the subsequent updates to this code, such as: r1337403, r1337475, r1337736, r1338059, r1338338. Votes: +0: rhuijben (I would like to see some test results before backporting) * r1409804 Add a delimiter between the two chains in svn_error_compose_create(). Justification: Allow users to tell where the second error chain starts -- avoids misrepresentation that the first error is caused by the second. Votes: -0: rhuijben (I'm not sure if this should be in a patch release and the stacktrace where this refers to is maintainer only) +1: danielsh (err->file and err->line are only in maintainer builds. this is about err->message which is valid in all builds) * r1421541, r1424977 Fix infinite recursion in './get-deps.sh'. Justification: Trivial fix to plug an edge case. Votes: -0: rhuijben (Is this needed on this branch?) +1: danielsh (as much as it's needed on trunk) * r1423837, r1423840, r1423848 Fix issue #4168, "no such table: revert_list" for nested wc in sparse wc Justification: Accessing the wrong database from the revert code gives an ugly error which might not be easy to work around. Votes: +1: rhuijben, pburba * r1426752 Fix Issue #4133, an issue in the diff-libraries whitespace normalization handling that might make diff use the wrong tokens, showing up as unexpected changes. Justification: The diff library should behave properly when suppressing whitespace changes. Notes: The branch is just needed for avoiding a conflict with the testcase. Will probably conflict again if r1426762 and/or r1427278 are merged to 1.7.x, because they all add a test to the testlist in diff-diff3-test.c. Branch: ^/subversion/branches/1.7.x-r1426752 Votes: +1: rhuijben, jcorvel * r1426762 Fix issue #4283, When the identical suffix starts at chunk boundary, "svn diff" produces incorrect diff. Justification: Without this patch the identical suffix might show up as an added or deleted suffix of a file. Notes: Will probably conflict if r1426752 and/or r1427278 are merged to 1.7.x, because they all add a test to the testlist in diff-diff3-test.c. Votes: +1: rhuijben, jcorvel * r1426830 Resolve issue #3797, by making sure that local changes at the same path as a not present file are not skipped. Justification: Improves 'svn diff' handling of local changes. Votes: +1: rhuijben, pburba * r1427278 Fix another issue in the diff-libraries whitespace normalization handling that might make diff show unexpected changes, and even trying to read beyond end of file. Justification: The diff library should behave properly when suppressing whitespace changes. See also http://svn.haxx.se/dev/archive-2012-12/0556.shtml for some examples of this bug. Notes: The branch is just needed for avoiding a conflict with adding a comment. Will probably conflict again if r1426752 and/or r1426762 are merged to 1.7.x, because they all add a test to the testlist in diff-diff3-test.c. Branch: ^/subversion/branches/1.7.x-r1427278 Votes: +1: jcorvel, rhuijben * r1427197, r1427210 Fix issue #4270 'diff -cN shows wrong line changed'. In the optimized code path (unaligned access) for prefix/suffix scanning, the diff would be incorrect when a '\r' is not immediately followed by a '\n'. Justification: Diff should show the correct changes. Failure to do so can lead to incorrect updates/merges (because diff3 suffers from the same bug). Branch: ^/subversion/branches/1.7.x-issue4270 Votes: +1: jcorvel, rhuijben * r1434405, r1434414, r1434418 Fix issue #4294 'diff --git shows wrong copy from'. Moving a dir causes diff --git to output incorrect sources for the diff. Justification: Diff should show the correct output. Failiure to do so can lead to incorrect interoperation with other systems. Votes: +1: breser +1: rhuijben (even though it only fixes one of the many --git issues) * r1434128, r1434435, r1434476 Fix issue #4034, 'blame on files with non-native eol-style sees every line as modified'. Justification: When blaming to the working copy instead of the base version eol styles all other styles than those with "\n" repository form show up as entirely changed. Votes: +1: rhuijben * r1434750 Resolve regression in 'svn status -u' output when running on externals. Without this patch the paths in directory externals are shown as absolute while all others are shown relative/as provided by the user. Justification: Consistent api behavior is goog Votes: +1: rhuijben * r1435635, r1435714, r1435722, r1435737 When svn update removes a switched subtree because it was deleted in the target revision, it would be nice if a future update would bring back the originally node that existed before switching. (Issue #4295) Justification: After this operation the working copy doesn't match the repository without obviouse way to recover. Notes: The real fix is in r1435722. The other revisions tweak the test that also triggered a different revision on trunk. Branch: ^/subversion/branches/1.7.x-issue-4295 Votes: +1: rhuijben * r1438602, r1438683, r1441810 Fix issue #4306, "multiple editor drive file merges record wrong mergeinfo during conflicts" Justification: Bogus mergeinfo is created during conflicting merges with file targets. Notes: r1438602 is the initial test and r1438683 the initial fix. r1441810 is a follow-up that addressed a bug in the initial fix (see http://svn.haxx.se/dev/archive-2013-02/0001.shtml) and also reworked the test to use prop edits rather than text edits as the changes to be merged. Unfortunately, 1.7.x has unrelated problems with property conflicts: It doesn't abort the merge if it encounters them! It merrily plugs along sans any "One or more conflicts were produced while merging...resolve all conflicts and rerun the merge to apply the remaining unmerged revisions" error (this is fixed on trunk). So while r1441810 is a valid fix (IMO) its test fails for unrelated reasons. So I reworked the earlier version of test to demonstrate the problem r1441810 fixes, but using file conflicts. Branch: ^/subversion/branches/1.7.x-issue4306 Votes: +1: stsp (r1438602 and r1438683 only), pburba * r1438879 Make 'show-diff' at the interactive conflict prompt show nicer diffs. Justification: 'show-diff' produces unreadable diffs in some cases, especially with conflicts recorded during merges. Branch: ^/subversion/branches/1.7.x-r1438879 Votes: +1: stsp * r1445753 Fix issue #3741, 'externals not removed when working copy is made shallow'. Justification: Externals linger after update --set-depth empty, causing unnecessary tree conflicts when the working copy is updated back to depth infinity. See regression test and also the recipe script attached to issue #3741. Votes: +1: stsp, rhuijben * r1451678, r1452617 Make get-deps.sh compatible with Solaris /bin/sh. Justification: Make get-deps.sh work on Solaris. Votes: +1: breser +0: rhuijben (untested) * r1452780 Address issue #4331, "working copy permissions change on commit for files with keywords". Justification: Keyword translation should not affect file permission bits. Votes: +1: stsp +0: rhuijben (code looks ok, but is a no-op on Windows) * ^/subversion/branches/1.7.x-gssapi-solaris10 Justification: Make it easier to build svn+serf with kerberos support on Solaris10. Branch: ^/subversion/branches/1.7.x-gssapi-solaris10 Notes: rhuijben: Is this necessary for 1.7. I think gssapi was moved to serf with/before serf 1.0, which is a 1.7 requirement? Votes: +1: breser * r1461562, r1461580 Fix issue #4340, "fs layer should reject filenames with trailing \n" Justification: Control characters can severely corrupt FSFS revision files and should never enter the repository for this reason. See discussion linked form issue #4340 for more information. Branch: ^/subversion/branches/1.7.x-issue4340 Votes: +1: stsp -1: cmpilato ("How did this expand from 'newlines' to 'control characters'?!) * r1461743 svn_fs_commit_txn: set output parameter in non-maintainer-with-special-flags builds.c Justification: Part of the API contract. Branch: 1.7.x-r1461743 Votes: +1: danielsh Veto-blocked changes: ===================== Approved changes: =================