cURL / Mailing Lists / curl-library / Single Mail

curl-library

test 56 on IRIX 6.2 MIPS C 6.2 autobuilds

From: Yang Tse <yangsita_at_gmail.com>
Date: Sat, 21 Apr 2007 12:36:21 +0200

Hi Daniel et all,

Failure of test #56 on IRIX 6.2 MIPS C 6.2 autobuilds has always been
somewhat random. The failure could appear or disappear when some new
code was commited to CVS. This has always puzzled me, since many times
the code changed didn't seem to have much to do with that specific
test case.

I think, the last episode of this test case # 56 appearing and
disappearing has to do with the changes that have been done to
Curl_pgrsUpdate. When I modified it test 56 seemed to work again on
mentioned platform. The changes introduced didn't intend to fix this
test failure, only intended to allow calling more frequently the user
supplied progress callback.

But my change introduced an undesired effect, the internal progress
meter would also be called more than once a second. Daniel detected my
error and fixed it
http://cool.haxx.se/cvs.cgi/curl/lib/progress.c.diff?r1=1.83&r2=1.84

Now the funny thing is that test #56 fails again on both IRIX 6.2
MIPS C 6.2 autobuilds.

Daniel's fix is of course completely right. But I don't understand
what kind of lurking bug is somewhere that only affects this platform
and that for example gets covered or uncovered by the mentioned
change.

Perhaps the more or less frequent flushing of the output stream has
something to do with it? Lurking bug actually in the library or in the
testsuite server?

Just wanted to share this out loud in case someone had any clue.

Or maybe Tor could locally verify if this change actually un/covers
the bug or if it makes no difference and test #56 on that platform
keeps being a mystery.

-- 
-=[Yang]=-
Received on 2007-04-21