[00:04] evand: I could squeeze in a basic workaround for half of bug 445067, perhaps [00:04] Launchpad bug 445067 in ubiquity "ubiquity overwrites VBR of extended partition" [Undecided,New] https://launchpad.net/bugs/445067 [00:05] ah, nice [00:08] partman-basicmethods: cjwatson * r762 ubuntu/ (choose_method/dont_use/do_option debian/changelog): [00:08] partman-basicmethods: Don't send the SET_FLAGS command when the new flags are the same as the [00:08] partman-basicmethods: old ones; this avoids partition tables being rewritten when all we did [00:08] partman-basicmethods: was set the method to "don't use" (LP: #445067). [00:20] partman-basicmethods: cjwatson * r763 ubuntu/debian/control: Maintainer and Vcs-Bzr for Ubuntu [00:26] partman-basicmethods: cjwatson * r764 ubuntu/debian/changelog: releasing version 43ubuntu1 [00:40] cjwatson: obviously not for karmic, but does something like http://pastebin.ubuntu.com/300153/ seem reasonable to handle IO errors occurring outside of the file copy routine that are not handled (such as in bug 445681)? [00:40] Launchpad bug 445681 in ubiquity "Kubuntu 9.10 beta: installer crashes before finishing install" [Undecided,New] https://launchpad.net/bugs/445681 [00:41] that is, simply moving the EnvironmentError exception handler one level higher [00:47] ENOENT might be a bit too general at that level, and quite possibly ENOTDIR too, but the others would be OK to move up like that, I think [00:47] okay, thanks [00:49] ubiquity: cjwatson * r3558 ubiquity/ (debian/changelog scripts/install.py): [00:49] ubiquity: Furthermore, always consider English as "complete enough". The packages [00:49] ubiquity: that are missing from an installation from the Ubuntu desktop CD are not [00:49] ubiquity: critical for a reasonable user experience. [01:00] cjwatson: Ah. I like r3558 -- that had me really confused. [01:00] Thanks. [01:06] no problem, not sure if the same needs to be done for pkgsel [01:06] I don't think I noticed it on an alternate install but I'll probably have to check again [01:07] * cjwatson falls over [01:08] heh [01:08] oh wow, it's 1am [02:19] so I would like to know more about what the GUI installer is doing which takes such an exorbitant amount of time when custom-editing the partitions ? with several disks, it can take minutes of scanning [02:19] and why there is no option to do a straight partition editor like in other distros [02:20] once modifying any partition, it proceeds to do the lengthy scan again [02:20] and again after another change. it is mind-bogglingly slow and frustrating to install Ubuntu because of this, especially with 4+ disks [10:46] evand, cjwatson there is a new version for patch #452348 [12:20] Samus_Aran: I'd like to optimise it, certainly; I know exactly what's happening and why, although a full explanation is a bit more than I want to go into on a Saturday morning, especially when you present it in such a confrontational way ;-) [12:21] Samus_Aran: in brief it's due to details of the way code is reused from the alternate/server installer; we most certainly do not want to have to maintain two partitioning programs, so I would far prefer to address this by optimisation rather than by rewriting, just haven't had the time yet [12:21] Samus_Aran: the answers you got on #ubuntu-devel (probing for other OSes, etc.) were not well-informed [12:23] cjwatson: so what would the summary be of what it's doing, if not probing for OSes ? [12:24] the way in which partman (d-i's partitioner) works is in some ways not a brilliant match for ubiquity. In order to reuse its behaviour, ubiquity needs to effectively drive partman back and forward under the covers in order to construct a useful graphical view [12:24] like I say, I already consider this an important bug that needs to be fixed [12:25] so you aren't going to make me consider it any more important by haranguing us ;-) [12:25] apologies for this "Moonraker12" character on #ubuntu-devel, no idea who he/she is but he/she was out of line [12:26] in fact, optimising the partitioner is already one of the things I've noted as a project I'd like to take on for the lucid cycle [12:27] what language is the partitioner programmed in ? [12:27] a mix of shell, C, and python [12:27] what's the majority share ? [12:28] I'm not sure of the answer, and I'm not sure that the answer would be helpful anyway :-) [12:28] there's C to interface to libparted; shell to do most of the business logic; and python to wire it up to ubiquity [12:29] any meaningful work on this will involve all three [12:29] well I am an expert at coding shell script, and can only read Python and C. if there is any optimisation to do in the shell script side of things, I could help out [12:29] I appreciate the offer, and you're welcome to look into this yourself if you wish, although there is a substantial learning curve [12:30] as I said, it is something I plan to tackle over the next cycle, assuming my manager thinks it's worth my time :) [12:30] you're incorrect that no caching happens, BTW, it's just not adequate caching ... [12:31] basically when ubiquity asks partman to perform an action, right now it doesn't always have a way to tell which partitions are going to be affected by that, and has to rescan rather too much as a result [12:32] I can think of some possible ways to remedy that [12:32] anyway, have to go out [14:55] ubiquity: evand * r3559 ubiquity/ (d-i/manifest debian/changelog): [14:55] ubiquity: Automatic update of included source packages: grub-installer [14:55] ubiquity: 1.43ubuntu8, partman-basicmethods 43ubuntu1. [15:15] ubiquity: evand * r3560 ubiquity/debian/changelog: releasing version 2.0.4 [15:54] * evand blinks at bug 459550 [15:54] Launchpad bug 459550 in ubiquity "ubi-language failed with exit code 141" [Undecided,New] https://launchpad.net/bugs/459550 [21:49] evand: may I bother you a bit? This is the update I want to test https://launchpad.net/ubuntu/+source/partman-basicmethods/43ubuntu1 unfortuantely I get stuck at the very beginning of your suggestion :( What is anna-install? from where should it be executed? (google ain't my friend). [23:09] arand_: I thought your problem was in the desktop installer, not in the text installer? [23:09] if so, anna-install is no help [23:10] arand_: if I'm right, open a terminal and run 'sudo apt-get update; sudo apt-get install ubiquity-frontend-gtk' (which will upgrade the installer) and then start the installer [23:16] arand_: oh, not until it gets processed through the queue though ... [23:18] arand_: how about: sudo apt-get install patch patchutils; wget -O- http://bazaar.launchpad.net/%7Eubuntu-core-dev/partman-basicmethods/ubuntu/diff/762 | filterdiff -x debian/changelog | sudo patch /lib/partman/choose_method/70dont_use/do_option [23:18] arand_: (I think - that's untested) [23:45] arand_: switch to a virtual terminal (ctrl-alt-f2) [23:46] then hit enter and type anna-install openssh-client-udeb [23:47] actually [23:47] you don't need to do that [23:47] still switch to the virtual terminal [23:47] 23:09 arand_: I thought your problem was in the desktop installer, not in the text installer? [23:47] and wget http://launchpadlibrarian.net/34315822/partman-basicmethods_43ubuntu1_all.udeb [23:47] 23:09 if so, anna-install is no help [23:47] ah [23:47] 23:10 arand_: if I'm right, open a terminal and run 'sudo apt-get update; sudo apt-get install ubiquity-frontend-gtk' (which will upgrade the installer) and then start the installer [23:47] my mistake for assuming [23:47] 23:16 arand_: oh, not until it gets processed through the queue though ... [23:47] 23:18 arand_: how about: sudo apt-get install patch patchutils; wget -O- http://bazaar.launchpad.net/%7Eubuntu-core-dev/partman-basicmethods/ubuntu/diff/762 | filterdiff -x debian/changelog | sudo [23:47] patch /lib/partman/choose_method/70dont_use/do_option [23:47] 23:18 arand_: (I think - that's untested) [23:48] though carry on in case I'm wrong :-) [23:48] udpkg -i partman-basicmethods_43ubuntu1_all.udeb