It seems than when you are working on a SVN branch, the SVN revision pointed by HEAD will *not* be the last change done on your branch, but the last change done on trunk (which is irrelevant for your checkout) See this example: 1. Check out webkitgtk-2.28.0 svn branch $ svn co http://svn.webkit.org/repository/webkit/releases/WebKitGTK/webkit-2.28.0 build # .... lot of files Checked out revision 266187. 2. Check what svn info says: $ cd build $ svn info Path: . Working Copy Root Path: /tmp/build URL: http://svn.webkit.org/repository/webkit/releases/WebKitGTK/webkit-2.28.0 Relative URL: ^/releases/WebKitGTK/webkit-2.28.0 Repository Root: http://svn.webkit.org/repository/webkit Repository UUID: 268f45cc-cd09-0410-ab3c-d52691b4dbfc Revision: 266187 Node Kind: directory Schedule: normal Last Changed Author: carlosgc@webkit.org Last Changed Rev: 258193 Last Changed Date: 2020-03-10 12:10:24 +0100 (Tue, 10 Mar 2020) So, the last change done on that branch was in reality r258193 (2020-03-10), but svn is saying that my working copy is r266187 (basically trunk of today) This causes the tool glib-revision to print the wrong revision, which then is passed to the build and the version stamped for the build is incorrect. $ Tools/glib/svn-revision r266187 I think we should pick the "Last Changed Rev" changeset instead of the "Revision" changeset from svn info.
Created attachment 407331 [details] Patch
Comment on attachment 407331 [details] Patch Unfortunate we have no tests for this.....
(In reply to Jonathan Bedard from comment #2) > Comment on attachment 407331 [details] > Patch > > Unfortunate we have no tests for this..... (To be clear, this is not a request to add tests, simply a comment on how we've organized this part of webkitpy)
Yes, i also wished to had some tests. I did manual tests as much as possible and it seems to work as expected. Thanks for the review ! :)
Committed r266267: <https://trac.webkit.org/changeset/266267> All reviewed patches have been landed. Closing bug and clearing flags on attachment 407331 [details].
<rdar://problem/67911890>