On 16/05/07, Karl Hasselström wrote: > On 2007-05-15 21:01:43 +0100, Catalin Marinas wrote: > > > What is the impact on the bash completion for calling StGIT rather > > than reading those files? Is it visible? > > Yes, it's visible, but not annoying (to me anyway). The overhead is > akin to the overhead we used to have when "stg help" generated the > command names -- on the order of 100-200 ms, when StGIT is in the > cache. The expensive part is to start stgit; the git calls are cheap. > So theoretically the completion script could duplicate the logic in > StGIT and avoid most of the overhead, if someone wanted it badly > enough. I did a quick test of 'stg series' with the DAG patches applied, on a Linux kernel repository ('du -sh .git' is 285M) with 42 patches (only 25 applied). It constantly takes over 2 seconds to complete (compared to < 200ms without the DAG patches). The problem is that this delay will happen for bash completion as well. It seems that most of the time is spent in git._output_lines() called from stack.read_refs() (for git-show-ref). I attach the profiling output generated by stg-prof. -- Catalin