Comment 5 for bug 670159

Hi all,
So, to be fair, I've sort of done an end-run on this one. I was looking for a unified diff parser for an independent project, and it just so happened bzr had one (sidenote: it would be cool if some of the awesome stuff in bzrlib was distributed separately!), so I was directly accessing the methods inside. The "bug" is not something I found from running bzr, and may never actually expose itself during normal operation.

Having read Aaron's reply, I went back and looked more closely at what I was doing, and he is right that the patch I fed the library was malformed. It was missing trailing lines, which I had accidentally stripped in my own pre-processing. It's your call whether you want patches.py to be more tolerant or not! I am happy with the "Invalid" resolution, and I'm sorry to have wasted time from user error :(

FWIW, the diff I was using (with the extra not-stripped newlines...) can be called up by running `svn diff -c 6977 https://svn.forge.morfeo-project.org/svn/libresoft-tools`