From: Borut R. <bor...@si...> - 2007-03-20 16:43:25
|
I would be much more satisfied it the pic16 reg. tests would complete without errors... My opinion is that reg. tests should show the problems, not hide them, but it seems that I'm in minority, so I: - disabled pic16 reg. tests on cf-x86 - corrected MacOS X 10.2 to MacOS X 10.4 - removed alpha, sun and openpower builds except the last one Borut Maarten Brock wrote: > I agree with Frieder. If nothing is broken in the stable targets the > indicator should be green. > > I see some of the old december files have already been removed. I wouldn't > mind to see it cleaned up some more: linux on alpha and openpower can do > with one entry too. > > And finally I was wondering if the OSX description is still correct. It > says MacOS X 10.2. Isn't that 10.4? > > Maarten > > >> Hi, >> >> I think it would make sense to momentarily remove >> pic16 from the nightly snapshot regression test >> of the compile farm. Arguments would be: >> >> The "single bit" information: "pic16 regression tests >> do not pass" currently would mask regressions for the >> other ports. >> >> The information "pic16 regression tests do not pass" >> is not hidden, as it is doc'ed f.e. here: >> http://sdcc.sourceforge.net/doc/sdccman.html/node130.html >> >> The "regression tests suite" is not used as a >> _regression_ test by pic16 (in the sense of ensuring >> that a previously fixed bug does not reappear again). >> While noone forces us to stick to the meaning of >> "regression" (and we could put "known to fail tests" >> there too) it probably makes sense now. >> Otherwise we'd have to include "known to fail tests" like: >> http://sourceforge.net/tracker/index.php?func=detail&aid=1536762&group_id=599&atid=100599 >> http://sourceforge.net/tracker/index.php?func=detail&aid=1618050&group_id=599&atid=100599 >> and the red/green dots on the snapshot page would >> permanently suggest the information "snapshot-20061231.xx.yy >> is better than "snapshot-today.xx.yy". (More generally >> "snapshot-old.xx.yy" better than "snapshot-now.xx.yy") >> >> >> Nicer than removing pic16 from nightly snapshot regression >> tests would of course be to exclude the (small fraction of!) >> problematic tests and warn somewhere about their exclusion. >> This would make sense for newly emerging bugs as well >> (kind of regression test pre-stage) but we currently do not >> have a mechanism for that? >> >> Greetings, >> >> Frieder >> >> ------------------------------------------------------------------------- >> Take Surveys. Earn Cash. Influence the Future of IT >> Join SourceForge.net's Techsay panel and you'll get the chance to share >> your >> opinions on IT & business topics through brief surveys-and earn cash >> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV >> _______________________________________________ >> sdcc-devel mailing list >> sdc...@li... >> https://lists.sourceforge.net/lists/listinfo/sdcc-devel >> >> >> > > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys-and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > sdcc-devel mailing list > sdc...@li... > https://lists.sourceforge.net/lists/listinfo/sdcc-devel > > |