Re: RFH: testsuite Heisenbugs

[Thorsten Glaser]

Dixi quod…

>Now, I’m suspecting a timing issue or something, so if anyone

… but while writing that, I saw this:

>       Argument [34] Oct 1999 [0-9][0-9]:00:00 -0000

>       Argument 3 Oct 1999 23:59:38 -0000

That made me curious, and indeed:

address@hidden:~ $ date; TZ=UTC0 date; TZ= date
Sat Oct 22 18:50:58 UTC 2016
Sat Oct 22 18:50:58 UTC 2016
Sat Oct 22 18:51:24 UTC 2016

So sanity.sh (which, IMHO, probably ought to have been called
insanity.sh instead ☺) fails to properly restore an unset $TZ.

Fix attached for your convenience.

PASS: test-getdate.sh
==================
All 1 tests passed
==================
Making check in diff
Making check in src
/usr/bin/make  check-local
/bin/mksh ./sanity.sh `pwd`/cvs
This test should produce no other output than this message, and a final "OK".
(Note that the test can take an hour or more to run and periodically stops
for as long as one minute.  Do not assume there is a problem just because
nothing seems to happen for a long time.  If you cannot live without
running status, try the command: `tail -f check.log' from another window.)
OK, all 3587 tests passed (11 test groups skipped).
/bin/mksh ./sanity.sh -r `pwd`/cvs
This test should produce no other output than this message, and a final "OK".
(Note that the test can take an hour or more to run and periodically stops
for as long as one minute.  Do not assume there is a problem just because
nothing seems to happen for a long time.  If you cannot live without
running status, try the command: `tail -f check.log' from another window.)
OK, all 3780 tests passed (1 test group skipped).
/bin/mksh ./sanity.sh -p `pwd`/cvs
This test should produce no other output than this message, and a final "OK".
(Note that the test can take an hour or more to run and periodically stops
for as long as one minute.  Do not assume there is a problem just because
nothing seems to happen for a long time.  If you cannot live without
running status, try the command: `tail -f check.log' from another window.)
OK, all 3614 tests passed (3 test groups skipped).

bye,
//mirabilos
-- 
“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
        -- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

sanity-fix.diff
Description: Text document