Mail Archives: cygwin-developers/1999/12/06/14:19:37
On Mon, Dec 06, 1999 at 01:08:13PM -0600, Mumit Khan wrote:
>Earnie Boyd <earnie_boyd AT yahoo DOT com> writes:
>>Since this changes the "stableness" of the snapshots, IMHO a copy of
>>the snapshots prior to these changes should be saved similar in nature
>>to the Jan 15th snapshot. BTW, I'm glad to see that pthreads will
>>finally get tested.
>
>I know that enabling threads breaks some of my code (** see below), so
>I agree with Earnie that we should have
>stable-before-the-last-major-change one. Perhaps 1999-12-04?
We provided the 1/15/99 version because there were problems with the B20.1
release. That's the only reason. I don't see this as a similar situation
at all.
The standard caveats apply. If enabling thread safety causes code to
hang then you can either try to fix the bug or build your own
non-thread-safe version.
I guess I could be convinced otherwise, but I don't view the snapshots
as a method for trying nifty new features while avoiding the responsibility
of testing.
The latest snapshot will actually fix a couple of bugs that are in the 12/4
snapshot. I don't want to be in the position of hearing about people with
bugs who are afraid to "upgrade" from 12/4 because cygwin is now thread-safe.
Of course, there is plenty of time to make this decision. If I start seeing
bug reports, I might reconsider and make a 12/4 snapshot specially available.
I certainly hope that it doesn't come to this.
cgf
>** It has been a while since I tried running a thread-enabled version,
>so chances are that a lot of the "usual" thread related problems have
>been solved. I was getting random hangs here and there among other
>problems. This code needs platforms that support DCE or POSIX (pthreads
>now works under Mingw), and currently runs under many different OSs;
>however, that doesn't mean that there aren't bugs in it.
- Raw text -