* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * THIS RELEASE STREAM IS OPEN TO BUG FIXES. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * This file tracks the status of releases in the 1.5.x line. See "Stabilizing and maintaining releases" in trunk/www/hacking.html for details on how release lines and voting work, what kinds of bugs can delay a release, etc. Candidate changes for 1.5.5: ============================ * r30257, r33024, r33026 Don't try to filter natural history from added mergeinfo during reverse merges. Notes: r30257 is the urelated removal of an unused private struct in merge.c, not required but avoids a conflict with r33024. r33024 is a change to include more information in the merge command baton, it results in no functional change itself but r33026 relies on this new information. There's a single trailing whitespace conflict during merge of r33024. r33026 is the core change. Justification: A reverse merge should simply return mergeinfo to a previous representation -- see http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=142777. This change also improves reverse merge performance as it eliminates some unnecessary communication with the repository. Votes: +1: pburba * r33866 Make ra_svn's Cyrus auth implementation (like the internal auth implementation) always prefer EXTERNAL to ANONYMOUS. This prevents a weird regression vs 1.4 where svn+ssh:// would be treated as anonymous in some cases. Justification: Regression against 1.4 which can prevent users from even checking out repositories that they could check out with a 1.4 client. Votes: +1: glasser * r34082, r34096 Fix issue #3275 by making WebDAV proxy slave servers pass through any GET or PROPFIND requests aimed at working resource URLs, even though those requests are read-only. Why? It does no good for the slave server to try to answer a read request aimed at an uncommitted transaction when said transaction's data lives solely on the master server. Votes: +1: cmpilato * r31311, r31314 Fix bug whereby 'svn cp --parents URL/TO/EXISTING/FILE URL/TO/EXISTING/DIR' would seg fault. Notes: r31311 is the testcase and r31314 the fix. Backport branch available as ^/branches/1.5.x-r31314. Justification: I just lost a night tracking this issue for an AnkhSVN user and I saw a few similar error reports earlier of users that couldn't reproduce. Votes: +1: rhuijben, hwright * r34385, r34393 Fix broken merge when the merge target's natural history includes resurrections. Notes: r34385 is a new test. r34393 is the fix. Justification: This problem was encountered 'in the wild' - see http://svn.haxx.se/dev/archive-2008-11/0618.shtml. When encountered cherry harvest merges are completely broken. Votes: +1: pburba * r34423 Fix issue #3312 "Mergeinfo for '...' maps to an empty revision range" due to sparse merging. Justification: Some people do merge into sparse directories or targets with switched subtrees; this produces non-inheritable mergeinfo. This non-inheritable mergeinfo can cause an error during merging or log -g. See issue #3312 and http://svn.haxx.se/dev/archive-2008-10/1170.shtml. Votes: +1: pburba, rhuijben * r34432 Follow-up to r32975, which is already backported and had a bug, see http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=145757. Notes: There is a very minor whitespace conflict in merge_props_changed(). Justification: r32975 fixed a bug which occurred when a merge added explicit mergeinfo to a subtree that had no explicit mergeinfo prior to the merge. But this fix itself had a bug which could result in an error during a --dry-run merge, r34432 fixes this. Votes: +1: pburba Approved changes: =================