=== robbiew is now known as robbiew-afk
=== robbiew is now known as robbiew_
=== saispo_ is now known as saispo
[09:28] <CIA-4> debian-installer: cjwatson * r1217 ubuntu/debian/changelog: No-change rebuild to pick up new components.
[09:31] <CIA-4> debian-installer: cjwatson * r1218 ubuntu/debian/changelog: releasing version 20081029ubuntu75
[12:31] <ev> I nearly have keyboard detection working in ubiquity.  Though I still need to figure out how to properly map kernel keycodes to X keycodes, rather than just adding 8 to the kernel keycode (evdev with no further mapping).
[12:33] <ev> Hrm, I just realized that foundations-lucid-oem-config-localized-keyboard-layout is not actually on the schedule.
[12:34] <ev> cjwatson: am I correct in assuming it's too late to approve it so that it appears on the work items page?
[12:35] <cjwatson> ev: it'll make the graph suck :) but I think it's possible
[12:35] <cjwatson> please get it fully drafted though
[12:35] <ev> sure thing
[12:36] <ev> on that now
[12:36] <cjwatson> three work items or so shouldn't break the bank, but are you happy with that on your own schedule?
[12:37] <ev> yes
[12:43] <cjwatson> kernel keycodes to X keycodes> doesn't console-setup do that?
[12:56] <ev> perhaps it does.  I haven't dug deep enough into its code.  I've been looking at code around xkb.
[12:58] <cjwatson> ev: the alternative is that keymapper is supposed to be able to deal with constructing maps based on pure X11 symbols
[12:58] <cjwatson> I'm sure it's bitrotted or something since we don't use it, but that might be easier than mapping back and forward
[12:59] <ev> ah, indeed.  I thought it might be able to.  I'll look into that.  Thanks!
[13:24] <CIA-4> ubiquity: cjwatson * r3623 ubiquity/ (bin/ubiquity debian/changelog):
[13:24] <CIA-4> ubiquity: Set sensible permissions on /var/log/installer/debug, not os.open's
[13:24] <CIA-4> ubiquity: default of 0755.
[13:28] <ev> drafting> done
[13:58] <CIA-4> partman-base: cjwatson * r176 ubuntu/ (debian/changelog lib/base.sh):
[13:58] <CIA-4> partman-base: Call sed outside debconf_select's inner loop. In my benchmarks using two
[13:58] <CIA-4> partman-base: disks with eight partitions each, this reduces debconf_select's runtime
[13:58] <CIA-4> partman-base: on partman/choose_partition from 0.69 seconds to 0.07 seconds.
[13:59] <ev> nice
[14:02] <cjwatson> total "Scanning disks" benchmark runtime goes from 74.7 seconds to 52.6 seconds. Still a long way to go.
[14:06] <cjwatson> BTW, I ended up not using bootchart; it produced huge image files and wasn't really quite as informative as I needed anyway. I've ended up just using a little bit of instrumentation in /lib/partman/lib/base.sh, ubiquity/debconffilter.py, and ubiquity/filtercommand.py instead.
[14:07] <cjwatson> and I just run with 'ubiquity -d', enter the manual partitioner, copy out partman and debug, and compare timings from the second occurrence of "Scanning disks" to the end of the file
[14:32] <CIA-4> partman-base: cjwatson * r177 ubuntu/ (debian/changelog init.d/update_partitions lib/base.sh):
[14:32] <CIA-4> partman-base: Cache the output of partition_tree_choices for each disk, invalidating
[14:32] <CIA-4> partman-base: the cache whenever we update a partition on the disk. In the above
[14:32] <CIA-4> partman-base: benchmark, this saves on the order of half a second every time we
[14:32] <CIA-4> partman-base: redisplay the partition tree when nothing has changed (e.g. on backing
[14:32] <CIA-4> partman-base: up from a partition).
[14:47] <CIA-4> ubiquity: cjwatson * r3624 ubiquity/ (debian/changelog ubiquity/components/partman.py):
[14:47] <CIA-4> ubiquity: Gather basic information from parted_server using the PARTITIONS command
[14:47] <CIA-4> ubiquity: (all partitions on a disk at once), which is slightly faster than
[14:47] <CIA-4> ubiquity: PARTITION_INFO (one partition at a time).
[14:49] <ev> noted, thanks
[14:54] <CIA-4> ubiquity: cjwatson * r3625 ubiquity/ (debian/changelog ubiquity/frontend/base.py):
[14:54] <CIA-4> ubiquity: Initialise automation_error_cmd, error_cmd, and success_cmd to '' rather
[14:54] <CIA-4> ubiquity: than None, since that's what's checked by the functions that use them.
[16:03] <cjwatson> ubiquity isn't happy with cdebconf right now; can't quite work out why
[16:12] <davmor2> cjwatson: ubiquity found out about cdebconfs affair with the alternate cd?
[16:13] <cjwatson> heh
[16:17] <CIA-4> ubiquity: cjwatson * r3626 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py):
[16:17] <CIA-4> ubiquity: Unlink /var/lib/localechooser/langlevel even if unlinking
[16:17] <CIA-4> ubiquity: /var/lib/localechooser/preseeded fails.
[16:36] <davmor2> cjwatson: the code for grabbing the lang packs is it hardcoded to where it grabs them from do you know off hand?
[16:37] <cjwatson> not offhand but I'd guess it just uses the configured mirror in /etc/apt/sources.list
[16:37] <davmor2> ta
[16:47] <CIA-4> ubiquity: cjwatson * r3627 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py):
[16:47] <CIA-4> ubiquity: Regain privileges to unlink /var/lib/localechooser/preseeded and
[16:47] <CIA-4> ubiquity: /var/lib/localechooser/langlevel.
[17:42] <CIA-4> ubiquity: cjwatson * r3628 ubiquity/ (bin/ubiquity debian/changelog ubiquity/debconffilter.py):
[17:42] <CIA-4> ubiquity: Enable debconffilter debugging if UBIQUITY_DEBUG_CORE=1, so that we can
[17:42] <CIA-4> ubiquity: set DEBCONF_DEBUG=developer rather than DEBCONF_DEBUG=developer|filter
[17:42] <CIA-4> ubiquity: which cdebconf doesn't understand.
[18:11] <cjwatson> dear cdebconf, why art thou made of abysmal
[18:12] <cjwatson> I'm pretty sure it just isn't coping with >1024-char input lines!
[18:45] <ev> lol
=== robbiew is now known as robbiew_
[22:26] <soren> cjwatson: Any clue how this could have happened: http://people.canonical.com/~soren/install_weirdness/
[22:27] <soren> cjwatson: (The weirdness being the title of the dialog box)
[22:40] <soren> Not exactly the kind of stuff I expected to catch with my automated testing :-/
[22:41] <cjwatson> soren: is it reproducible?
[22:42] <soren> Not sure.
[22:42] <soren> cjwatson: I'll check tomorrow. I need to head to bed now.
[22:42] <cjwatson> soren: if it is, I'd like to get a log with DEBCONF_DEBUG=developer on the boot line
[22:43] <soren> cjwatson: I'll look into that. It's running now, I'll check the results tomorrow morning.
[22:43] <cjwatson> ok, thanks
[22:43]  * soren disappears
[23:34] <CIA-4> partman-base: cjwatson * r178 ubuntu/debian/changelog: releasing version 135ubuntu2
=== mpt_ is now known as mpt
[23:51] <CIA-4> cdebconf: cjwatson * r1440 ubuntu/ (debian/changelog src/debconf-communicate.c):
[23:51] <CIA-4> cdebconf: * Backport from trunk:
[23:51] <CIA-4> cdebconf:  - Teach debconf-communicate to handle long input lines (over 1024
[23:51] <CIA-4> cdebconf:  bytes).
[23:51] <CIA-4> cdebconf: cjwatson * r1441 ubuntu/debian/control: Maintainer and Vcs-Bzr for Ubuntu
[23:53] <CIA-4> cdebconf: cjwatson * r1442 ubuntu/debian/changelog: releasing version 0.145ubuntu1
=== robbiew is now known as robbiew-afk
=== robbiew-afk is now known as robbiew_