Re: [PATCH] Speed up bash completion loading

> Stephen Boyd wrote:
> 
> > I see a small problem, but it can be fixed in another patch. git
> > merge -s <TAB><TAB> the first time when you're not in a git
> > directory will make git merge -s <TAB><TAB> after never complete
> > correctly even in a git directory.
> 
> Not good.  I think the sanest thing to do is avoid caching the merge
> strategy list entirely.  Listing merge strategies takes about 100 ms
> here, which is short enough not to be annoying.
> 
> > I guess this become more serious
> > if git isn't in your path initially and you do git <TAB><TAB> and
> > then git becomes part of your path and you try again. Here you lose
> > porcelain completion. This is probably never going to happen though,
> > right?
> 
> Right, if this happened to me I would not be too surprised.  A similar
> problem occurs when someone installs a new git subcommand in the
> middle of a session.  Starting a new session fixes the completion, but
> should the completion script do something to help people before then?
> 
> It could provide a shell function with a slightly friendlier name than
> "__git_compute_porcelain_commands" for the user to invoke explicitly.
> 
> It could retry "git help -a" each time completion was needed if it
> gave no results last time (i.e. use "${__git_all_commands:=" instead
> of "${__git_all_commands=").  This would help with the missing git
> problem (which seems unusual), but not the missing git-svn.
> 
> Maybe it could detect signs of user frustration (repeated attempts to
> complete the same thing?) and add to the frustration by silently
> fixing the cache.