Re: c++/544: gdb.c++/annota2.exp: annotate-quit test sometimes fails

[Michael Elizabeth Chastain <mec at shout dot net>]

The following reply was made to PR cli/544; it has been noted by GNATS.

From: Michael Elizabeth Chastain <mec@shout.net>
To: carlton@math.stanford.edu, gdb@sources.redhat.com
Cc: gdb-gnats@sources.redhat.com
Subject: Re: c++/544: gdb.c++/annota2.exp: annotate-quit test sometimes fails
Date: Fri, 7 Feb 2003 09:39:57 -0600

 I ran annota2.exp 10 times with my usual testbed for a total of 380 test
 runs.  annotate-quit had 380 PASSes and 0 non-PASSes.  This is for gdb
 HEAD date '2003-02-06 19:11:07 UTC'.
 
 The last report had gdb HEAD date '2003-02-05 07:30:23 UTC' and it had
 5 PASS, 33 KFAIL for this test.  This has been typical for years.
 
 (By the way, those dates are exact arguments to "cvs -D '...' checkout".
 I used to do "cvs checkout" and record the date in an XML file.
 Now I look at my clock (which is NTP sync'ed), subtract a couple
 of minutes, and do "cvs -D '...'" on that date.  So the date should
 be exactly reproducible for stuff like "cvs diff -D ... -D ...".
 Thanks to Andrew C for that tip).
 
 I diff'ed 2003-02-05 07:30:23 UTC and 2003-02-06 19:11:07 UTC.  The only
 patch of interest is the big Ingham/Seitz/Zannoni/Cagney interpreter
 change.  Fred Fish checked in a precedence fix to infttrace.c but the
 only effect of that fix is to change some gdb debugging messages.
 
 The cool part is that we will leave all the code in gdb.c++/annota2.exp
 exactly the same, so that we will get KFAIL's again if the bug comes
 back, and the KFAIL's will point to the PR, and the PR has this thread
 in the audit trail.  We can edit the comments in annota2.exp to reflect
 our new knowledge.
 
 I would like to put the PR into 'feedback' state for 30 days, and then
 close it.
 
 Michael C