> This type of bug is going to bite us over and over.
Does that mean we make every object threadsafe? IMHO, No.
Oh, I agree for sure.
Better that we document which returned apr_foo_t's are threadsafe, which ones are not,
Yep.
Nah. Subversion was doing something incredibly stupid. We were walking up the pools until it found the global_pool and then created the apr_xlate_t in there and just reused that. The global_pool has no guarantee on being thread-safe within the context of httpd.
It has nothing to do with apr_xlate_t. It has everything to do with using the global_pool and the false assumption that it was threadsafe.
I don't think it's a big deal that we (in APR) have to worry about, but that's just me. -- justin
