[02:21] <twb> Hello.  I just got a new laptop prototype unit to test oem installation for, and it turns out that the framebuffer during install is misaligned -- basically I get the left half of the display on the right side of the screen and vice-versa.
[02:21] <twb> I've never seen this before; any idea what's going on?
[07:01] <CIA-4> ubiquity: superm1 * r2383 ubiquity/debian/ (changelog ubiquity-frontend-mythbuntu.templates): correct small typo in mythbuntu template
[14:25] <cr3> when netinstalling ubuntu, is there a way to prevent the installer from downloading Packages.gz from security.ubuntu.com? the reason I think this is happening is that when I specify my own repository, a mounted alternate image, I get erros that some package like e2fsprogs is not available but it's only because the image has a slightly older version
[14:27] <cjwatson> I saw your question by mail, but have not worked out the answer yet; I will send it when I have done so
[14:48] <cr3> cjwatson: come to think of it, perhaps my network should be configured so that packets cannot go out
[14:49] <cjwatson> honestly I suspect you will just need to use a matching image
[15:21] <tjaalton> cjwatson: so, livecd-rootfs does a 'rm -f ${ROOT}etc/X11/xorg.conf', but could it create an empty file after that?
[15:27] <cr3> cjwatson: in the case of -proposed mentionned by pitty, this might pose a problem
[16:04] <cjwatson> cr3: you could use the server CD image built against -proposed
[16:04] <cjwatson> tjaalton: it can do whatever you like ;-)
[16:05] <cjwatson> tjaalton: feel free to change it in bzr and upload
[16:07] <cr3> cjwatson: right, I mount that CD on httpd, copy over the netinstall files to tftpd, boot the remote machines and it complains that some packages are not there. the Packages.gz file and pool directory on the CD are correct, so I suspect the netinstall is fetching some information from remote. is that possible?
[16:29] <cjwatson> I'll reply by mail
[17:21] <xivulon> cjwatson, the other day I stumbled upon the python module "subprocess", never used it myself, but thought it might have been useful with respect to what we discussed a few days ago'
[17:21] <cjwatson> xivulon: subprocess fails to handle the problem I described
[17:22] <cjwatson> unless you take special measures
[17:22] <cjwatson> I'm already using subprocess
[17:22] <xivulon> ok, didn't know whether you had tried that already or where using os.popen
[17:22] <cjwatson> to clarify, I'm not looking for help; I have addressed the problem to my satisfaction in my own code
[17:22] <cjwatson> that doesn't mean I don't think it's a bug in python with serious implications for using it as a shell replacement
[17:25] <cjwatson> bug 40464 is where I ran into it
[17:25] <ubotu> Launchpad bug 40464 in ubiquity "espresso crashes on partitioning step in Kubuntu 6.06 LTS Beta Live CD" [Critical,Fix released] https://launchpad.net/bugs/40464
[17:26] <xivulon> They are in the midst of a major rewrite for python 3k, if an appropriate pep is prepared, they might put that in
[17:27] <xivulon> 40464 must have been a fun one...
[17:28] <cjwatson> http://www.mailinglistarchive.com/python-dev@python.org/msg14630.html
[17:34] <xivulon> not much followup, hmm that's a shame
[17:39] <cjwatson> http://mail.python.org/pipermail/python-dev/2007-July/073831.html is a better URL, but still no followup
[17:42] <xivulon> I noticed there were no replies to the message, but there are a couple of open bug report that look quite similar: http://bugs.python.org/issue1615376
[17:43] <xivulon> http://bugs.python.org/issue1488934
[17:44] <cjwatson> 1615376 seems to be describing something similar but subtly different (which fds are open, not the signal mask)
[17:44] <cjwatson> though the reporter may be confused; but the bug is certainly not clear
[17:45] <xivulon> http://bugs.python.org/issue1068268
[17:45] <cjwatson> 1488934 may well be similar but it has not been properly analysed, and what the user is suggesting is actually doing the *opposite*
[17:46] <cjwatson> 1068268 is unrelated
[17:46] <xivulon> might be worth to do a proper bug report on the issue
[17:46] <cjwatson> I may do at some point
[17:54] <superm1> what should the appropriate behavior be if someone was to 're-use' a user name when preserving /home?  Should the uid be updated to be 1000:1000 throughout their home directory?
[17:54] <cjwatson> we put this in the spec
[17:55] <cjwatson> "#
[17:55] <cjwatson> "If the user created during the install already exists in /home, reuse the UID and GID from /home for that user by preseeding them.
[17:55] <cjwatson> "
[17:55] <cjwatson> i.e. change our UID/GID, not the filesystem
[17:55] <superm1> ah okay.  i didn't recall what was decided and have the spec handy
[17:55] <cjwatson> https://wiki.ubuntu.com/UbiquityPreserveHome
[18:12] <superm1> looking at user-setup source, it doesn't look like its been implemented yet, but the preseeded value for the gid i'm assuming will be in passwd/user-gid, correct?
[18:17] <xivulon> talking about users, where do I find a list of invalid usernames?
[18:17] <superm1> in the user-setup source
[18:18] <superm1> there is a file called reserved-usernames
[18:18] <xivulon> ah thanks
[18:41] <CIA-4> ubiquity: superm1 * r2385 ubiquity/ (debian/changelog scripts/mythbuntu/apply-type): check for existing files related to mythbuntu autostart in a user's home directory
[18:51] <cjwatson> superm1: yeah, I think adduser defaults to the same gid as the uid, but if they happen to differ on the existing filesystem then that would certainly need a user-setup extension
[18:51] <superm1> cjwatson, yeah my thoughts were if someone had installed on top of a different distro that didn't default that same way
[18:52] <cjwatson> please do send a patch for user-setup
[18:59] <superm1> cjwatson, is there a place you'd particularly like me to commit the patch to?
[18:59] <superm1> i only see a svn imported bzr branch in LP
[19:01] <superm1> oh nvm, its part of a core-dev branch.  i'll publish it to one of my own branches
[19:08] <superm1> cjwatson, okay I added the gid patch if you would like to merge my branch here: https://code.edge.launchpad.net/~superm1/user-setup/preservehome
[19:33] <tjaalton> cjwatson: heh, ok. I was just wondering if it was the right place to work around it :) I'll do it tomorrow
[19:34] <xivulon> cjwatson re wubi related bugs/patches, https://bugs.launchpad.net/wubi/+bugs?field.tag=wubi let me know if there is anything else I can do
[20:41] <CIA-4> ubiquity: superm1 * r2386 ubiquity/ (3 files in 3 dirs): Don't allow the user to progress on non mythbuntu master backend roles unless they test their connection
[22:11] <CIA-4> ubiquity: superm1 * r2388 ubiquity/scripts/mythbuntu/mythbuntu_install.py: archive uses dvb-utils not dvbutils