Re: PATCH: completion caching layer

[Adam Spiers <adam@spiers.net>]

Bart Schaefer (schaefer@candle.brasslantern.com) wrote:
> On Aug 1,  5:31pm, Adam Spiers wrote:
> } Subject: PATCH: completion caching layer
> }
> } +findex(_retrieve_cache)
> } +item(tt(_retrieve_cache) var(cache_identifier))(
> } +This function retrieves completion information from the file given by
> } +var(cache_identifier), stored in a directory specified by the
> } +tt(cache-path) style (defaults to tt(~/.zsh/cache)).
> 
> I'd prefer that the cache location be relative to $ZDOTDIR rather than
> relative to $HOME, and that it not be two levels removed, e.g. use
> $ZDODTIR/.zcompcache/ rather than ~/.zsh/cache/.  That'd be consistent
> with e.g. .zcompdump.

Sure, change applied, so will appear in the commit.  Makes no
difference to me; I already had

  zstyle ':completion::complete:*' cache-path ~/.zsh/cache/$HOST

in my .zshrc because my home directory is mounted via NFS so I need a
different cache path for each machine (for perldoc and rpm at least).

> } +  # grr, doesn't work, so we have to roll our own
> } +#  typeset "$@[2,-1]" > "$_cache_dir/$_cache_ident"
> 
> Isn't
> 
>     typeset +g "$@[2,-1]" > "$_cache_dir/$_cache_ident"

Ah, +g!  I was looking for that.

> what you want?  Or maybe
> 
>     while (($# > 1)); do
>       shift
>       ${(P)=1:+typeset $1}
>     done > "$_cache_dir/$_cache_ident"
> 
> would be more accurate?

Is that better than what you suggested via private mail?