gdb.c++/annota2.exp: annotate-quit

gdb.c++/annota2.exp: annotate-quit

[David Carlton]

Sometime between yesterday and today, the test gdb.c++/annota2.exp,
which had been consistently KFAILing for me, started consistently
passing.  Is anybody else seeing this?  Does anybody currently see it
KFAIL?  Is there any chance that the interpreter patch from yesterday
or today might have fixed this bug?  (The bug has nothing to do with
C++ or annotations: it's just that GDB wasn't always responding to a
^C as it should.)

David Carlton
carlton@math.stanford.edu

Re: gdb.c++/annota2.exp: annotate-quit

[Daniel Jacobowitz]

On Thu, Feb 06, 2003 at 12:39:23PM -0800, David Carlton wrote:
> Sometime between yesterday and today, the test gdb.c++/annota2.exp,
> which had been consistently KFAILing for me, started consistently
> passing.  Is anybody else seeing this?  Does anybody currently see it
> KFAIL?  Is there any chance that the interpreter patch from yesterday
> or today might have fixed this bug?  (The bug has nothing to do with
> C++ or annotations: it's just that GDB wasn't always responding to a
> ^C as it should.)

I see this also; but it's done this in the past, so I'm reluctant to
claim that it's actually fixed.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: gdb.c++/annota2.exp: annotate-quit

[Mark Kettenis]

David Carlton <carlton@math.stanford.edu> writes:

> Sometime between yesterday and today, the test gdb.c++/annota2.exp,
> which had been consistently KFAILing for me, started consistently
> passing.  Is anybody else seeing this?

Yes I do see this on i386-unknown-freebsd4.7.  My history shows that
the test has been consistently PASSing until january 5 2003 (at least
since august 12 2002), started KFAILing between january 5 and january
15, and started PASSing again somewhere between february 2 and
february 5.

Mark

Re: gdb.c++/annota2.exp: annotate-quit

[Michael Elizabeth Chastain]

dc> Is anybody else seeing this?
dc> Does anybody currently see it KFAIL?

I'll run it a couple of hundred times in the morning and report back.

dc> Is there any chance that the interpreter patch from yesterday
dc> or today might have fixed this bug?

Yes.  The test case reads a prompt from gdb, sends a ^C to gdb,
and expects gdb to react to it.  I suspect there's a race condition
involving job control so that the incoming ^C races with the
operating system enabling gdb to receive signals (as opposed to
the inferior receiving the signal).  So something that touches
the main prompt/read loop could touch this.

(I admit I'm guessing a lot here).

Michael C