On Thu, 2016-04-14 at 09:56 -0700, Junio C Hamano wrote: > David Turner <dturner@xxxxxxxxxxxxxxxx> writes: > > > $GIT_DIR/index-helper.path is a symlink to the socket for the > > daemon > > process. The daemon reads from the socket and executes commands. > > We generally avoid using symbolic links inside $GIT_DIR (we used to > represent the current branch by a symbolic link HEAD pointing at the > underlying ref, but switched to the textual symref long time ago, > and we did the ".git pointing at the real repository location" to > support submodules also in a similar way). > > I wonder if we can change this to a file that records the location > of the socket without bending over backwards too much? > > > Named pipes were considered for portability reasons, but then > > commands > > that need replies from the daemon would have open their own pipes, > > since a named pipe should only have one reader. Unix domain > > sockets > > don't have this problem. > > > > On webkit.git with index format v2, duplicating 8 times to 1.4m > > entries and 200MB in size: > > > > (vanilla) 0.986986364 s: read_index_from .git/index > > (index-helper) 0.267850279 s: read_index_from .git/index > > > > Interestingly with index v4, we get less out of index-helper. It > > makes > > sense as v4 requires more processing after loading the index: > > > > (vanilla) 0.722496666 s: read_index_from .git/index > > (index-helper) 0.302741500 s: read_index_from .git/index > > > > (these benchmarks are from an earlier version of this patch, but > > should > > still be valid). > > Just something to keep in mind; we'd need to see updated numbers > when the series matures. I don't actually know how Duy did the index duplication for this benchmark. Duy? -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html