On 08/01/14 21:13, David Kastrup wrote:
Actually, that seems like a mischaracterization to me. It's the single file logs which aren't cheaper than the multi-file logs.
You could be right. In any case, the Facebook devs are claiming wins for all change-related commands, including status, diff, update and commit:
https://code.facebook.com/posts/218678814984400/scaling-mercurial-at-facebook/
At the current point of time, due to a discussion about version control on the Emacs developer list, I am working on git blame which is ridiculously slow. For the file src/xdisp.c, it takes about 2 minutes on my computer. I've cut it down by some 20 or 30% on the first attempt, but with my current approach, I should eventually get it down to few seconds at most. At the current point of time, it segfaults almost instantaneously. One out of two ain't bad... Ridiculous that this code has not been sensibly improved over the last 10 years or so...
It's a surprise to me, certainly. What do you think of the Facebook developers' contention that Git's internals would be difficult to work with if serious scaling gains are desired?
_______________________________________________ lilypond-user mailing list lilypond-user@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-user
