This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: increasing default timeout significantly
- From: "Carlos O'Donell" <carlos at redhat dot com>
- To: Andreas Schwab <schwab at linux-m68k dot org>, libc-alpha at sourceware dot org
- Date: Wed, 20 Jan 2016 00:33:25 -0500
- Subject: Re: increasing default timeout significantly
- Authentication-results: sourceware.org; auth=none
- References: <20160119211840 dot GJ14840 at vapier dot lan> <87r3hdxm2r dot fsf at igel dot home> <20160119222646 dot GL14840 at vapier dot lan> <87mvs1xk72 dot fsf at igel dot home> <20160119225028 dot GN14840 at vapier dot lan>
On 01/19/2016 05:50 PM, Mike Frysinger wrote:
> On 19 Jan 2016 23:45, Andreas Schwab wrote:
>> Mike Frysinger <vapier@gentoo.org> writes:
>>> in what way ? there are really only three states:
>>> - test doesn't set a timeout
>>> - a passing test doesn't care if you timeout in 2 sec or 1 year
>>> - a failing test would take longer to timeout now, but that test
>>> should already be logged, and i don't think this should affect
>>> the default to the detriment of the common case & valid systems
>>> - test sets a higher timeout
>>> - the custom timeout is still used
>>> - behavior is unchanged
>>> - test sets a lower timeout
>>> - the custom timeout is still used
>>> - behavior is unchanged
>>>
>>> now, if you mean "you should delete the #define TIMEOUT from all tests
>>> whose value is <=20", then yes, that's a cleanup that i'd probably do.
>>> but i wasn't going to bother updating >=50 files if we didn't want to
>>> accept the fundamental change i posted above.
>>
>> That should have been part of the original submission.
>
> i disagree. if people don't think the default TIMEOUT should be
> increased, updating 50+ files is a complete waste of my time. i
> don't like wasting my time..
I think updating TIMEOUT in all tests that define TIMEOUT is an orthogonal
cleanup that should not block increasing the default timeout for tests that
don't define TIMEOUT.
To put it another way: I think Mike's change to increase timeout to 20 is
fine. In practice I run with TIMEOUTFACTOR=999999 on my development system
to catch hung tests and debug them.
Cheers,
Carlos.