=== greenwhich_ is now known as greenwhich
=== davmor2 is now known as davmor2_dinner
=== davmor2_dinner is now known as davmor2
[14:08] <CIA-1> debian-installer: cjwatson * r914 hardy-proposed/ (9 files in 4 dirs): Move to 2.6.24-17 kernels.
[14:16] <evand> re 224446> sounds reasonable (and yikes bzr is going slow today).
[14:18] <cjwatson> I'm concerned that there is no ubiquity upload to hardy-proposed yet
[14:18] <cjwatson> do you need help?
[14:22] <evand> I should be ok, I'm just dealing with bzr weirdness at the moment, then I'll take care of uploading it.
[14:22] <cjwatson> ok
[15:09] <evand> cjwatson: Should I add hardy-proposed to the hardy-proposed version of ubiquity's sources.list?
[15:09] <cjwatson> I think so
[15:10] <cjwatson> I think there are a few differences
[15:12] <evand> ok
[15:20] <CIA-1> debian-installer: cjwatson * r915 hardy-proposed/ (build/Makefile debian/changelog):
[15:20] <CIA-1> debian-installer: * Reinstate Fabio's changes from dapper-updates:
[15:20] <CIA-1> debian-installer:  - Build against non-release pockets matching the sources.list on the
[15:20] <CIA-1> debian-installer:  buildd.
[15:20] <CIA-1> debian-installer:  - Dump the generated sources.list.udeb to the build log.
[15:25] <CIA-1> debian-installer: cjwatson * r914 ubuntu/ (5 files in 5 dirs):
[15:25] <CIA-1> debian-installer: Bump hd-media image size to 800MB so that Ubuntu ISO images fit on them
[15:25] <CIA-1> debian-installer: (LP: #225402).
[15:26] <CIA-1> debian-installer: cjwatson * r916 hardy-proposed/debian/changelog: version number for hardy-proposed
[15:27] <CIA-1> debian-installer: cjwatson * r917 hardy-proposed/ (5 files in 5 dirs): merge from trunk
[15:30] <evand> hrm, "nominate for release" on LP seems to be broken.
[15:32] <cjwatson> has been working for me - what's up with it?
[15:34] <evand> weird, nevermind it's working now.  One of the bugs I was working with would go to a 404-type page when I clicked on that option but I can't reproduce it now.
[15:35] <cjwatson> are you going to include a translations update
[15:35] <cjwatson> ?
[15:35] <cjwatson> (I think you should) if so, watch out for the German one - I included a manual fix there for a frequently-reported mistake on the front page, which may need to be reapplied
[15:36] <evand> ok, will do
[15:36] <saispo> cjwatson: http://people.ubuntu.com/~ubuntu-archive/seeds/ubuntu.hardy/ is up2date ?
[15:36] <cjwatson> I imagine so, why?
[15:37] <cjwatson> it's updated by a cron job and generally left along
[15:37] <cjwatson> alone
[15:37] <saispo> because in STRUCTURE you use an "include platform.hardy"
[15:37] <cjwatson> though it's better to use the ones on bazaar.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/ubuntu.hardy/ if possible
[15:37] <saispo> but no see where is it :)
[15:37] <cjwatson> yes, that's correct
[15:37] <cjwatson> /seeds/platform.hardy/
[15:37] <cjwatson> it's including a whole other seed collection
[15:38] <saispo> where can i found it ?
[15:38] <cjwatson> use your imagination with what I just said and the URL you started with :)
[15:38] <saispo> thks :)
[15:38] <saispo> i will not use the first i posted :)
[15:38] <cjwatson> it should work fine ...
[15:39] <saispo> hmmm
[15:39] <saispo> will try
[15:39] <cjwatson> that's where germinate in hardy points by default
[15:39] <cjwatson> (unless you use --bzr in which case it goes to bazaar.launchpad.net)
[15:39] <saispo> yes but i include one file for my personnal packages
[15:40] <saispo> all work, thks cjwatson :)
[15:45] <CIA-1> debian-installer: cjwatson * r915 ubuntu/build/config/ (4 files in 4 dirs): do a better job of increasing hd-media to 800MB
[15:46] <CIA-1> debian-installer: cjwatson * r918 hardy-proposed/build/config/ (4 files in 4 dirs): merge from trunk
[15:49] <CIA-1> debian-installer: cjwatson * r919 hardy-proposed/debian/changelog: releasing version 20070308ubuntu40.1
[16:35] <evand> just waiting on the translations download link from LP.
[16:53] <evand> FWIW, the manual fix for de.po appears to no longer be necessary.
[17:53] <evand> ugh, I missed the RAW*SIZE change for ubiquity 1.8.8.
[18:25] <evand> https://bugs.launchpad.net/ubuntu/+bug/235374/+nominate - 404
[18:25] <evand> does it work for anyone else?
[18:46] <mario_limonciell> yeah i get a 404 too
[18:46] <mario_limonciell> try off edge maybe?
[18:51] <evand> no such luck
[19:18] <CIA-1> ubiquity: evand * r2685 ubiquity/ (configure configure.ac): Bump to 1.9.0
[19:23] <CIA-1> ubiquity: evand * r2686 ubiquity/debian/po/ (21 files): Update translations from Launchpad.
[19:36] <CIA-1> ubiquity: evand * r2687 ubiquity/debian/changelog: releasing version 1.9.0
[19:38] <CIA-1> ubiquity: evand * r2688 ubiquity/ (configure configure.ac): Bump to 1.9.1
[19:50] <mpt> evand, did you report that LP problem already?
[19:51] <evand> mpt: negative
[19:51] <evand> should I, or do you know if they're already aware of it?
[19:51] <mpt> evand, ah, never mind, it's reported as bug 176492
[19:52] <evand> ah, indeed
[20:05] <greearb> So, I fixed the generation of the md5sums.txt file so that it ignored some files that were transient...now my custom live CD passes the md5sum check.
[20:05] <greearb> but, it still fails to install..claiming that there is an IO error and the HD and/or CD is corrupt.
[20:06] <greearb> I tried two different machines (and different CDs), with same results.
[20:06] <greearb> is there any way to get an idea of exactly what file it is trying to muck with when the error occurs?
[20:14] <evand> greearb: it should SUBST in the filename in the question before it gets asked.
[20:14] <greearb> sorry, that made no sense to me :)
[20:15] <evand> oh, actually, nevermind I'm wrong.
[20:15] <greearb> ubiquity gets to about 80% of copying files, then it dies...I used a verbose flag to get a log file, but it doesn't show any file specifics
[20:16] <greearb> I assume some underlying tool is reporting errors...but not sure how to get at more specific error messages.
[20:19] <evand> It currently only reports the filename when the checksumming failed, not on any other file copy method.
[20:19] <evand> For the purposes of testing, you could modify /usr/share/ubiquity/scripts/install.py as such:
[20:19] <evand> http://paste.ubuntu.com/16460/
[20:20] <greearb> I don't think it's a checksumming problem...it seems more like when it's actually trying to copy files?
[20:20] <greearb> ok, I see
[20:20] <evand> greearb: I think I may have confused you.  I was saying that in terms of all of the copying errors ubiquity currently reports, it only displays the filename in the case of a MD5 mismatch on copy error
[20:21] <greearb> can I do this mod on a live cd after booting it, or do I have to respin the cd with this change?
[20:21] <evand> you can modify it in the live environment
[20:22] <greearb> ok, thanks..will give that a try right after lunch
[20:23] <evand> you're welcome
[22:05] <greearb> looks like /var/lib/scrollkeeper/TOC/1644 - 1651 are corrupted
[22:05] <greearb> IO errors trying to do much of anything with them
[22:06] <cjwatson> not an installer problem
[22:06] <cjwatson> oh, sorry
[22:07] <cjwatson> I completely ignored the context, so please ignore me ...
[22:07] <greearb> probably not :)
[22:07] <cjwatson> the md5sum check is not unfortunately 100% accurate
[22:07] <cjwatson> seems that normal use of the live CD tends to trip slightly different sets of problems
[22:07] <greearb> might consider adding that dbg patch pasted earlier when -d is used..very useful for these types of problems
[22:08] <greearb> maybe the md5 is just fine, but the image was packed up wrong?
[22:08] <greearb> ie, isofs or something else is funky?
[22:08] <greearb> I need to check my build image too..maybe it is funky
[22:09] <greearb> for that matter, have the installer print out the file name that failed
[22:21] <cjwatson> greearb: isofs> in theory possible, but staggeringly unlikely
[22:22] <cjwatson> greearb: if that were the case, everyone would have the same problem, and that isn't the case
[22:22] <cjwatson> and even discounting that, bear in mind that the md5 checker mounts the image and walks through it file by file checking md5sums
[22:22] <cjwatson> so it would surely suffer from the same problems
[22:23] <greearb> the md5 checker just seems to check the compressed file-system
[22:23] <greearb> (and some other files on teh CDROM)
[22:23] <cjwatson> right, you mean squashfs then rather than isofs
[22:23] <greearb> right, sorry about that
[22:23] <cjwatson> so that is certainly possible in theory, but it would be very odd for it to manifest on just a few files
[22:24] <greearb> yes, it is likely my bug for now...will know more when I get a chance to look at the source image on my HD.
[22:24] <cjwatson> maybe we should have the md5 checker mount the squashfs as well and walk through it
[22:24] <cjwatson> might stand more chance of exercising identical bugs
[22:25] <greearb> then you have to create md5sums differently..and many files can change legitimately, so that might be tricky to do right
[22:25] <greearb> maybe could somehow detect all files that came straight from a .deb and use .deb's md5sums (assuming it has such) to very only those files?
[22:25] <greearb> still could have false positives..but should ignore some obvious ones at least.
[22:28] <cjwatson> deb sort of has md5sums though only by policy not mandated by the format (so they aren't complete)
[22:29] <cjwatson> that would definitely be complicated for people who decide to make cowboy changes though, which is common enough (heck, done it myself)
[22:29] <cjwatson> probably best stick with what we've got
[22:30] <CIA-1> debian-installer: cjwatson * r920 hardy-proposed/ (34 files in 2 dirs): Update translations from Launchpad.
[22:30] <greearb> nod
[22:31] <greearb> only thing I'd ask for at this point is better error message letting user know what file failed and where source could not be read or dest could not be written.
[22:34] <CIA-1> debian-installer: cjwatson * r921 hardy-proposed/debian/changelog: releasing version 20070308ubuntu40.2
[22:59] <CIA-1> oem-config: cjwatson * r466 hardy-proposed/debian/ (58 files in 2 dirs): Update translations from Launchpad.
[23:01] <CIA-1> oem-config: cjwatson * r467 hardy-proposed/ (configure configure.ac): bump to 1.37.2
[23:02] <CIA-1> oem-config: cjwatson * r468 hardy-proposed/debian/changelog: releasing version 1.37.2
[23:43] <greearb> ok, so if I look at the uncompressed live-cd tree (ie, before squashfs is built, etc), then the scrollkeeper files look fine
[23:43] <greearb> but, if I mount the live-cd image generated from this, and mount the squashfs inside it, then I get the IO errors when looking at the file
[23:44] <greearb> so, it's not bad CD media, at least, as there is no CD involved at this point
[23:49] <greearb> looks like removing scrollkeeper removes the entire desktop :P
[23:53] <cjwatson> interesting
[23:53] <cjwatson> maybe you've hit another corner case in squashfs? we had one a while back that was basically dependent on image size ...
[23:59] <greearb> I'm not too sure how to debug that part much further..any ideas?