Re: svn commit: r16244 - in trunk: . subversion/clients/cmdline subversion/include subversion/libsvn_client subversion/libsvn_wc subversion/svnversion subversion/tests/clients/cmdline subversion/tests/clients/cmdline/svntest tools/examples

Garrett Rooney wrote:

>On 9/24/05, brane@tigris.org <brane@tigris.org> wrote:
>
>
>
>>* subversion/libsvn_subr/cmdline.c: Include apr_atomic.h and apr_pools.h.
>> (svn_apr_initialize): New. Initialization for apr and apr-atomic.
>> (svn_cmdline_init) Use svn_apr_initialize instead of apr_initialize.
>>
>>
>
>I'm not exactly clear why we need our own apr_initialize wrapper...
>
>All it seems to do is call apr_initialize then apr_atomic_init, but
>apr_atomic_init is already called when you call apr_initialize, at
>least on unix, it's not called in apr_initialize directly, but it's
>called as part of apr_pool_initialize, which apr_initialize calls, at
>least according to comments in apr/misc/unix/start.c. Is this
>different on windows or something? Even if it is, that seems like the
>kind of thing that should be fixed in APR, not in a Subversion
>specific wrapper.
>
>
Quite frankly, I didn't notice that. The APR documentation doesn't
promise to call apr_atomic_init -- although in retrospect, it seems
obvious that it should.

I'm all for throwing out that custom initializer and the relevant
docstring bit in svn_wc.h.

-- Brane

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Sep 25 04:42:32 2005