William A. Rowe, Jr. wrote:
>-1 (veto) to both suggestions.
>
>This is getting obscene. We've lost all value of tables, since they can't be
>memcpy'ed,
>they have to be reconstructed. This will be much slower than the original array code.
>
>Committers, please review dir_create() and dir_merge() patches more carefully. There
>is a long tradition of configuration bugs due to these assumptions. The patches that
>added copy/copy_mappings was an attempt to salvage this.
>
>I'll demonstrate a solution, using pool scopes, that adds safety and speed. Give me
>a day.
>
Are you still considering optimizing directory_walk so that it
doesn't have to do all the dir_merge work on each request (either
through a caching solution or through some variant of the pre-merge
patch that I posted a while back)?
If we didn't have to call merge_mime_dir_config with such alarming
frequency, it might not be necessary to optimize this function (or
dir_merge functions in general).
--Brian
