On Fri, Jul 08, 2005 at 04:57:18PM +0100, Stephen Stafford wrote: > Quoting Steve Langasek <[EMAIL PROTECTED]>:
> > I can't seem to reproduce this bug with an etch system (or at least, it > > takes a very long time testing mirrors -- hours -- before it segfaults). > > Is this bug actually confirmed? > Hi Steve, > I've confirmed it with sarge and sid. You can shorten the time it takes to > check mirrors with "-e 3" or similar. However you should see the segfault > before any mirrors are tested since it's happening during the parsing of the > mirror list. Did you "apt-spy update" before trying? Aha, I guess I missed the apt-spy update step. Yep, perfectly reproducible here now as well. > [EMAIL PROTECTED]:~$ apt-spy -d foo -o foo.list > Segmentation fault > > The problem appears to be that the format of the README.mirrors.txt file has > changed to add an "architectures" column. Unfortunately, due to a chronic > attack of Real Life[tm] I've not had a chance to chase it properly. > > We may have to change the format of the mirrors file back (which will 'fix' it > trivially), or keep the new format, but with a new name. Again this will > 'fix' > it trivially. > > I wasn't aware that the team responsible for that file were going to change > the > format, so I wasn't watching for this. Apologies for lack of vigilance. > > I'll do my best to find time to chase this and come up with a fix next week. I think there is other software in sarge that already uses this architecture field, unfortunately. Perhaps it would be easier to get an apt-spy update into sarge that supports the new format. Cheers, -- Steve Langasek postmodern programmer
signature.asc
Description: Digital signature
