apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject Re: documented 1.0 showstoppers
Date Fri, 04 Jun 2004 13:07:15 GMT
Joe,

  so you are suggesting since it's broke on Unix (but works on Win32
and possibly elsewhere), that nested|default|unnested will go away?

  In the meantime "The test suite is passing on most of the Unixes", 
and "People have lived with this 'broken' API forever" seem to be the
only justifications to release APR now, as-is.

  Please do not change the locking API at all.  Have unix return the
APR_ENOTIMPL to a nested lock create request, and rip out the code
we all concur is bogus from the unix side.  Pretty trivial really.

Bill

At 05:45 AM 6/4/2004, Joe Orton wrote:
>Agreed that none of those are really showstoppers. The test suite is
>passing on most of the Unixes I run nightly builds on; notable omission
>was FreeBSD because of the globalmutex thing, but I think Justin just
>may have just fixed that.
>
>The only other thing I can think of is the infernal nested mutex flag; I
>will just go ahead and remove the flag from the API but leave it for
>internal APR use since nobody else had a better idea on moving forward
>with that issue.
>
>Oh: what about the random code?  That got marked as a "1.1" feature in
>CHANGES by someone but the code is still built tested and invoked in
>HEAD...
>
>joe



Mime
View raw message