This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
[Bug translator/2156] check return value for _stp_pmap_agg()
- From: "joshua dot i dot stone at intel dot com" <sourceware-bugzilla at sourceware dot org>
- To: systemtap at sources dot redhat dot com
- Date: 18 Jan 2006 18:54:21 -0000
- Subject: [Bug translator/2156] check return value for _stp_pmap_agg()
- References: <20060116083509.2156.hunt@redhat.com>
- Reply-to: sourceware-bugzilla at sourceware dot org
------- Additional Comments From joshua dot i dot stone at intel dot com 2006-01-18 18:54 -------
(In reply to comment #4)
> > buildok/pmap_foreach.stp tests the unsorted case - I just modified it to test
> > the sorted case as well.
>
> OK, but that tests only buildability. Is there a run-time test?
I will make sure that there's a runtime test. If possible, I will also try to
create tests of the failure case as well.
(In reply to comment #5)
> _stp_pmap_agg() will return NULL if the map pointer is invalid, the map type in
> unknown, or it overflows. The first two should be impossible with code from the
> translator, so I suggest changing the error message to indicate an overflow
> happened and suggest the map size needs increased.
Ok, I will change the message as suggested.
> _stp_pmap_get_agg() always returns a valid pointer but in the case above the
> aggragated map will not be complete. So printing it anyway is safe, but
> potentially misleading.
There won't be anything misleading, because the first statement within the
foreach (maybe a print, or anything) will check c->last_error before doing
anything. Thus the error is delayed a little bit, but as long as the operations
between the error and its detection are safe, there's no problem. I see no
reason to optimize for the error case and try to catch it sooner.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=2156
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.