[Bug-readline] Re: [patch] Fix testsuite annotate-quit race (PR 544)

[Jan Kratochvil]

On Wed, 19 Mar 2008 10:55:39 +0100, Nick Roberts wrote:
> I am just worried it will break something for Emacs just as a previous change
> that you and Daniel made to readline did.

http://sourceware.org/ml/gdb-patches/2006-11/threads.html#00234
http://sourceware.org/ml/gdb-patches/2006-12/threads.html#00009


> It appear to me that the only problem this FAIL is causing is with the actual
> running of the testsuite.  I would be quite happy if this test was just
> removed.  Would that work for everyone else?

The problem is unrelated to the annotations and there should be a testcase just
with the "(gdb) " prompt after the annotations support gets obsoleted/removed.

Currently the test described at
        http://sourceware.org/ml/gdb-patches/2008-03/msg00270.html
has different output depending on whether you type CTRL-C early enough (during
the "(gdb) " printing) or just later (in poll(2) waiting for input).

With the readline patch proposed there the output is always the same.

I find it a bug as without the artifical `sleep (1)' delay it is very racy.


> but I am just worried it will break something for Emacs just as a previous
> change that you and Daniel made to readline did.

My original patch requiring a readline change AFAIK did not break anything,
unfortunately it did not get accepted for readline.  Daniel created an IMO
complicated patch to workaround the readline problem without forking readline
but unfortunately it brought that regression.  I find the regression was in
fact only a consequence of the readline interface deficiency.


Regards,
Jan