|
|
Well, that's a bunch for you guys to chew on. I'm done flooding for a
little bit now. For reference, that was 20 commits.
I did *not* develope these in order. All this code started out as a
single, frigging huge commit. Then, I uses git rebase -i, to turn it
into smaller and smaller commits. Then, I would compile stuff at each
commit; oops, doesn't compile, let's reorder the commits. Oh, wait,
here's a bug, let's move that commit earlier, just for the heck of it.
Ok, now let's go thru every commit, adding license headers. Ok,
another go around adding some javadocs. Ok, test cases. Oops,
discovered a bug in several earlier commits, fixed them. Bother, this
commit needs to actuall be earlier, because even tho the new feature
doesn't require it, the test case does. Another full commit history
edit, removing all debug stmts. Hmm, not getting full test coverage
in some places, let's fix each commit that has that issue.
Ok, one final loop over all 20 commits, testing each one in turn;
good, all done. Let's start a commit flood.
ps: that large commit is still sitting in my queue; it was actually
split up into 33 separate commits, I only sent the first 20. Still
writing more test cases.
pps: the actual real commons-vfs code has this information output:
git diff 40fc8dff7f702a1e1f5a0111d1449c9be304df46^
40fc8dff7f702a1e1f5a0111d1449c9be304df46|diffstat | tail -n 1
101 files changed, 8958 insertions(+)
(note, that there are no license headers on those 100+ files)
|
Locked
