Johan Corveleyn wrote on Tue, Sep 28, 2010 at 23:37:23 +0200: > On Tue, Sep 28, 2010 at 4:11 PM, Daniel Shahaf <d...@daniel.shahaf.name> > wrote: > >> Index: subversion/include/svn_diff.h > >> =================================================================== > >> --- subversion/include/svn_diff.h (revision 1001548) > >> +++ subversion/include/svn_diff.h (working copy) > >> @@ -112,6 +112,11 @@ > > (personally I prefer 'svn diff -x-p' to show the function name here) > > Ok, will do next time. >
Thanks. > I've had some progress: > - The blame_tests.py now all pass (tests 2 and 7 provoked a core > dump). That makes all tests pass. However, although I fixed the > coredump, I don't fully understand the root cause (why they ended up > in the situation where they ended up). So I'm going to study that > first a bit more. > - I have now included support for files with \r eol-style. > > I'll send a new version of the patch shortly. > Sounds good. > I'm also thinking that I could try to take advantage of -x-b/-x-w or > -x--ignore-eol-style options to make it even faster (right now the > prefix/suffix scanning will stop at the first difference, regardless > if it's a whitespace or eol difference that could/should be ignored). > ... > > Thoughts? None, actually; I'm not (yet) sufficiently familiar with diff internals. But let's please have this work in small, digestible (read: reviewable) pieces --- support for -x--foo is exactly the sort of additional optimization that can be done in a separate patch (on top of this starting patch).