On Wed, Jun 15, 2011 at 5:12 PM, Russell Dickenson <[email protected]> wrote: > On 16 June 2011 08:00, François BIOT <[email protected]> wrote: >> Hello >> I had this problem severall times. And it is very annoying. >> If suppose that the corruption is either a bug in pacman (too many >> corruptions in fact) or a lake of crc check. >> Pacman should either reject db with a crc or be smart enough to download the >> db when a package corruption is persistent. >> At work, other people than me hate frugalware when this occurs (and when I'm >> not there ;-)) >> <off-topic> can you imagine : I've decied to install many frugalware in my >> company so each time anything is bad : this is ''my distro'' ;-) >> ]</off-topic> >> Anyway, fixing this problem with a simple crc should be enough (if >> possible). >> Another idea should be tp 'mark' somewhere that the package nnn was >> corrupted and if it is one more time corrupted, ask the user to run -Syy or >> download automatically a new db :-) >> Regards, Frugalware rocks :-) >> [ and my game is still not finished, even if we can play with it ] >> 2011/6/14 Russell Dickenson <[email protected]> >>> >>> On 15 June 2011 06:29, Marius Cirsta <[email protected]> wrote: >>> > On Tue, Jun 14, 2011 at 10:53 PM, Melko <[email protected]> wrote: >>> >> 2011/6/14 Marius Cirsta <[email protected]> >>> >>> >>> >>> That fixed it indeed. It's still strange though because it did >>> >>> download it with pacman -S qt4-plugin-sqlite3 too. It asked if I >>> >>> wanted to delete it , I then said yes and it downloaded the file again >>> >>> and I did this several times. >>> >>> >>> >>> Might be a bug in pacaman-g2 or it might have been cached somewhere >>> >>> but I don't understand how... >>> >>> >>> >>> Thanks. >>> >>> Thanks anyway ... se >>> >> >>> >> The problem was that you had a local db version corrupted, pacman -Syy >>> >> forced to download the db also if it was up to date. >>> >> Melko >>> >> _______________________________________________ >>> >> Frugalware-devel mailing list >>> >> [email protected] >>> >> http://frugalware.org/mailman/listinfo/frugalware-devel >>> >> >>> >> >>> > >>> > Yes , that makes sense. Shouldn't this happen atumatically or >>> > something ? I mean if I were a regular user and hadn't asked here it >>> > really would have been a tough one to figure out.... >>> > >>> > Just a thought though ... >>> >>> I agree because I recently had exactly the same problem. It's not at >>> all obvious that the problem lies not with the mirror(s) but the local >>> package database. Is there some way this might be automatically >>> detected and a "pacman -Syy" be forced? >>> >>> >>> -- >>> May you always be Frugal, >>> >>> Russell Dickenson (AKA phayz) > > > I agree that if there is some way of preventing these errors being > presented to users then it should be done. I understand that someone > might change a package but forget to update the SHA1SUM. If this > happens though, perhaps pacman-g2 could force the "pacman-g2 -Syy" if > it detects this error the first time and if a second package fails its > checksum check then perhaps raise a fatal error and stop the > operation? > > > -- > May you always be Frugal, > > Russell Dickenson (AKA phayz) > _______________________________________________ > Frugalware-devel mailing list > [email protected] > http://frugalware.org/mailman/listinfo/frugalware-devel >
Wouldn't it be possible to somehow impose an increase of either the pkgver or pkgrel ? I mean if you do update a package you should probably increase one of these. Maybe the build machine could check this ? Or maybe on commit but I'm not sure .... _______________________________________________ Frugalware-devel mailing list [email protected] http://frugalware.org/mailman/listinfo/frugalware-devel
