proposal for permissions on mutexes

[Aaron Bannert <aa...@clove.org>]

Would something like this even make sense for non-Unix platforms?

apr_proc_mutex_chown(apr_uid_t owner, apr_gid_t group);
apr_proc_mutex_chmod(apr_mode_t mode);

(where apr_mode_t is the same thing as apr_fileperms_t but not
specifically for files, and takes the bitwise-OR of the
APR_UREAD-like symbols)

The same two functions would be useful for these types:

apr_proc_mutex_t
apr_global_mutex_t
apr_shm_t

-aaron

RE: proposal for permissions on mutexes

[Ryan Bloom <rb...@covalent.net>]

The apr_fileperms_t stuff is completely bogus for non-Unix platforms.
They were a shortcut, because abstracting that stuff out is really hard
to do.  I tried a few years ago, and ended up giving up.  We really need
to make a concerted effort to abstract out the permissions stuff, before
creating those functions.

Ryan

> From: Aaron Bannert [mailto:aaron@clove.org]
> 
> Would something like this even make sense for non-Unix platforms?
> 
> apr_proc_mutex_chown(apr_uid_t owner, apr_gid_t group);
> apr_proc_mutex_chmod(apr_mode_t mode);
> 
> (where apr_mode_t is the same thing as apr_fileperms_t but not
> specifically for files, and takes the bitwise-OR of the
> APR_UREAD-like symbols)
> 
> The same two functions would be useful for these types:
> 
> apr_proc_mutex_t
> apr_global_mutex_t
> apr_shm_t
> 
> -aaron