> Win32: apr_thread_rwlock_try*lock(), apr_thread_cond_timedwait(),
> apr_proc_mutex_*() (Is proc_mutex unnecessary on Win32?)
apr_thread_rwlock_try*lock() may now work on Win32. I (more or less) copied
the code from the OS/2 version. I didn't test it other than to run the
existing test suite on it. If the existing test suite is sufficient, it
works, otherwise it needs verification.
> * I think apr_open_stderr() and friends *should* dup() the
> descriptor. That would allow the new/returned file to be closed
> (via pool cleanup or manually) without accidentally closing
> stderr/out.
Hey, I noticed (crashed doing) that! I thought it was just supposed to work
that way. Now that you mention it, using dup() in this matter would be
better.
> Documentation that needs writing:
>
> * API documentation
The apr_vformatter documentation doesn't seem to have all of its escape
characters documented, only the %pp variation from the norm. I had to look
in the code for the 'q' length specification. Moreover, the documentation
that _is_ there is hard to read.
> Stuff for post 1.0:
I would suggest:
* Routines for acquiring basic information about running processes
such as memory usage, CPU usage, etc. Should work for processes
that were _not_ started via apr_proc_create(). Should allow
enumeration of all processes on system (subject to user permissions).
But that's just me, and probably no one else cares. ;)
mma
