[Monetdb-developers] MonetDB Testing & Status-Alert

Stefan Manegold Stefan.Manegold at cwi.nl
Sun Apr 25 00:35:00 CEST 2004


Dear fellow developers,

let me first apologize for the fact that I haven't been able to maintain the
nightly testing as usual while I was abroad the week before last week (Sa,
April 10 - Su, April 18). During this time, problems with SourceForge (that
are beyond our control) as well as problem with the Windows testing (that
are somehow under our control, see below), caused the automatic testing to
fail, and hence could not provide you with the usual feedback about your
checkins. 

Furthermore, I also have to apologize, that I did not manage to get the
testing working again, during last week, while being back at work.

Well, at least, the SF problems seems to be gone, again.

The windows problem is still there: recent changes in the release branch(!)
of MonetDB seem to cause "Application Errors" (aka SegFaults) on Windows,
and my non-interactive test-environment is unfortunately not able to press
the "Ok" button on the respective pop-up window...
Well, thanks to some hint from our new friends at SAP, I (hopefully) just
managed to avoid these pop-up windows on our Windows test machine. Hence,
testing should hang there anymore, even in case of "Application Errors".

These problem solved, I had a short(!) look at today's Stable testweb, at
what I saw did *scare* me a bit:


! The current "stable" version of MonetDB (4.3.16) seems to be pretty broken !


While many difference are "only" due to "harmless" changes in the output
(e.g., caused by changes in the behaviour/output of commands such as
"atoms()"), 
	there are some changes that indicate that "join" is broken !

Well, I guess, I'll have to spend tomorrow or Monday on looking at the
details and sorting out, what did go wrong with your checkins to the (no
longer) "stable" release branch...

Admittedly, I have to blame myself, again: Not only did I not keep the
automatic testing properly running during the last two week, but I also
forgot to give any guidelines, how to handle checkins to the "stable"
branch after the release had been made...

(No, I don't metion anymore, that Mtest might be run by hand on your
favorite development platform, at least before checking-in "bug-fixes" to
the "stable" branch...)

To see it positively, this observation confirms that we do not "waste" any
time or money by having one person spending a major part of his/her
(official) working hours on "quality control" --- I just forgot to find/name
a stand-in while being abroad myself...).

Good night,

Stefan

-- 
| Dr. Stefan Manegold | mailto:Stefan.Manegold at cwi.nl |
| CWI,  P.O.Box 94079 | http://www.cwi.nl/~manegold/  |
| 1090 GB Amsterdam   | Tel.: +31 (20) 592-4212       |
| The Netherlands     | Fax : +31 (20) 592-4312       |




More information about the developers-list mailing list