I, and I imagine others, prefer to have a source repository to code against. I hate saying "in file /scp/tickets.php" around line 26 remove blah and insert this line...". To me a diff against a known code base makes life so much easier. It is human readable and more importantly machine readable.

Take my bug http://osticket.com/forums/project.php?issueid=145(http://osticket.com/forums/project.php?issueid=145) ...when rc6 comes out then unless the bug has been merged into head then I'm going to have to repatch but if I have the patches as diffs then it is trivial from command line to repatch the new release in seconds without any further thought.

Look at this mod here, http://osticket.com/forums/showthread.php?t=2719(http://osticket.com/forums/showthread.php?t=2719) which looks very useful - I have no real idea what rc it was built against but I think it was rc4 and now rc5 is out then it should be easy for the mod provider to diff that and just provide 1 file not a long description.

For me, I'm looking at my sourcefilters ( as described here http://osticket.com/forums/showthread.php?t=2767(http://osticket.com/forums/showthread.php?t=2767) which would be fairly meaty chunk of code but if when rc6/rc7 etc comes out I have to re-port and that's pain.

Now I can easily create my local git from the source files but that divorces me from the rest of the world in that my published diffs would make no real sense to other people unless they then created their own git repository.

It's not like it costs money to run - there are a number of places that'll run free code repositories (sourceforge and github) - the issue is to be able to issue patches against known (e.g. head or branch tagged) code bases which everyone else understands what we're patching against.