Thomas Rast <trast@xxxxxxxxxxxxxxx> wrote: > However, I also wanted to provide tests, and that's where the fun > started. Turns out t9119-git-svn-info.sh is currently quite broken. > 1-4 just fix the tests; a brief summary: > > [1/6] git svn info: tests: let 'init' test run with SVN 1.5 > > The tests do not report any problems with 1.5. > > [2/6] git svn info: tests: do not use set -e > > No idea how 'set -e' ever got in there; it completely breaks the > test script in case of an error. I have a habit of using set -e in my scripts since I often forget (or am too lazy) to check for errors when executing a series of commands. > [4/6] git svn info: tests: fix ptouch argument order in setup > > Swapped arguments caused ptouch to fail and tests to break. Hm... I seem to remember explicitly setting the arguments one way for one reason or another. > Yes, I'm just whoring commit karma here, so feel free to squash these > four into one if you like it better that way. > > After these, 22 of 37 tests (all except --url, plus all unknown files) > fail. Most of them are caused by 'git svn info' not URL-encoding the > URL and Repository fields in the output, as SVN does. 6/6 fixes > this. > > However, the unknown files tests still fail, simply because 'svn info' > itself fails on untracked files. It would be great if someone who is > still running SVN 1.4 could check whether that has been a bug in > git-svn all along, or is actually a behaviour change on the part of > SVN. Oops, I upgraded to 1.5.x here already. I should still have another machine with 1.4 to check on tomorrow, though. -- Eric Wong -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
