meeting log for 12-16-2004

12:25 < codeman> i guess we should start this meeting.
12:25 < wolf31o2-work> fsck your meeting
12:25 < wolf31o2-work> *grin*
12:25 < samyron> lol
12:25 < z1nOnly|Work> agaffney: i am looking at the gtk partitioning screen.  in the alpha release, will you be able to select file type to be used?
12:26 < codeman> lets start w/ a status report on the GUIs (will give more time for esammer to get back)
12:26 < agaffney> z1nOnly|Work: you mean filesystem type?
12:26 < z1nOnly|Work> agaffney: yes, filesystem type
12:26 < AllanonJL|W> i work on all the screens except the 2 agaffney works on, ( partitioning and NFS mounts )
12:26 < AllanonJL|W> or panels if you want to call them that
12:27 < agaffney> z1nOnly|Work: yes, I just haven't put that stuff in yet
12:27 < wolf31o2-work> I guess I'm here as like... the releng representative or something
12:27 < samyron> hehe
12:27 < codeman> AllanonJL|W: have you gotten them to spit out partial XML profiles yet?
12:27 < agaffney> sure, why not
12:27 < AllanonJL|W> codeman: yes
12:27 < wolf31o2-work> will you be able to select reiser4?
12:27  * wolf31o2-work ducks
12:28  * agaffney stabs wolf31o2-work in the eye
12:28 < codeman> how close is it to being done (not including partitioning and NFS mounts)?
12:28 < agaffney> codeman: do you mean "working" or *done*?
12:28 < codeman> working, sorry
12:28 < agaffney> working...not long
12:28  * AllanonJL|W agrees with agaffney
12:29 < codeman> i remember seeing the partitioning screen going through  a variety of changes
12:29 < agaffney> and it will continue to develop
12:29 < codeman> i know its a rough screen for any installer
12:30 < BenUrban> heh, this isn't just any installer :P
12:30 < wolf31o2-work> just steal it from Red Hat... I mean... it can't be that hard, right?
12:30 < AllanonJL|W> codeman: i'd say to a close to full xml profile generation, probably 2 or 3 weeks into january it would be ready to test...it would be *rough* ( all the gui elements wouldn't be perfect according to HIG and other designs )
12:30  * agaffney stabs wolf31o2-work in the other eye
12:30 -!- spy|work [~donnie-VeFfXsABM5lHy2saJm6Ui8bjbc2VWiXoWuw85uim5zDR7s880joybQ@public.gmane.org] has joined #gentoo-installer
12:30 < samyron> wolf31o2-work: actually, stealing installer-stuff from redhat is a HUGE pain in the ass
12:31 < wolf31o2-work> I meant more the visual layout... not code
12:31 < samyron> o
12:31 < samyron> nm then
12:31 < codeman> the red hat installer has a lot more options
12:31 < agaffney> we aren't RH and we don't need to copy from them
12:31 < samyron> hrm.. we are gentoo.. we need more options than everyone
12:31 < samyron> lol
12:31 < spy|work> no, but we can learn from their UI.
12:31 < codeman> it can do RAID and all that jazz
12:32 < AllanonJL|W> also, who are we making this installer for? the average joe or a more technical user?
12:32 < samyron> AllanonJL: a little of column A and a little of column B
12:32 < codeman> both :)
12:32 < agaffney> AllanonJL|W: all of the above
12:32 < AllanonJL|W> :) ok
12:32  * BenUrban doesn't think that policy will last very long
12:32 < agaffney> for the first release, I'd rather cater to the automatic deployment crowd
12:33 < codeman> agaffney: your dialogfe, that greates a full xml profile?
12:33 < AllanonJL|W> wolf31o2-work: for the extra packages screen i was going to copy how redhat's install makes you select packages, are you familiar with that screen in the redhat install?
12:33 < samyron> yeah
12:33 < agaffney> codeman: almost a full profile. there were a few sections that weren't complete
12:33 < wolf31o2-work> AllanonJL: I am familiar with the RHEL stuff... not anything else
12:33 < samyron> just do it by category, like: net-www, media-libs?
12:33 < codeman> it would be really nice to have the dialog one done first so we can test it w/ integration and such
12:34  * agaffney pushes the dialogfe code in codeman's direction
12:34 < codeman> i was just gonna offer to do it too
12:34 < AllanonJL|W> oh yes, i remember, can we get a backend method for settings dhcpd options?
12:34 < wolf31o2-work> actually... the dialog one should be done first, from a releng standpoint, so we can get it on as "experimental" on the next release... I'm sure the gtk version will take much longer to perfect (and we'll ahve to be frozen somewhere around 2nd week of January for the next release)
12:34 < samyron> AllanonJL: yeah
12:34 < samyron> AllanonJL: we can
12:35 < codeman> AllanonJL|W: you never told us what options you wanted
12:35 < samyron> doesn't matter what options, i'll have a "generic" options field
12:35 < samyron> so if you specify something, it better work
12:35 < samyron> :-)
12:35  * BenUrban likes that philosophy
12:35 < AllanonJL|W> codeman: yeh i mentioned to agaffney, but that was the only one i came across so far that wasn't implemented in the backend :)
12:35 < codeman> there are still a few other functions not up to snuff
12:36 < codeman> the aliasing code in the networking is still commented out b/c it didn't work at all.
12:36 < agaffney> remember, this is only the *first* release
12:36 < codeman> true
12:36 < agaffney> we can add RAID, interface aliases, etc. in a future release
12:37 < AllanonJL|W> so timetable is 2nd week in january then? ( according to what wolf31o2-work said )
12:37 < samyron> uhm
12:37 < wolf31o2-work> can I just say something real quick
12:37 < samyron> aliases are no longer needed to be special
12:37 < samyron> because they are recognized by is_eth_device
12:37 < samyron> so you can speicfy an alias like a normal ethernet card
12:37 < samyron> at least in the Profile
12:37 < wolf31o2-work> the basic idea as I understand it is the installer needs to be able to do a "by the handbook" install for the first release... additional options can be added at a later time
12:38 < samyron> that's basically what we've been doing...
12:38 < wolf31o2-work> RAID/LVM/VLANS... all that sort of stuff can be added in later
12:38 < samyron> look at the client controller, that follows the handbook step-by-step
12:38 < wolf31o2-work> I know... I'm just saying... don't overwork yourselves
12:38 < AllanonJL|W> wolf31o2-work: okay, so i'll hold on the extra packages screen until after 2nd week in january ( because it will take awhile ) and just ask for a space seperated list in a box
12:39 < codeman> AllanonJL|W: sounds good.  functionality first. then looks.
12:39 < agaffney> AllanonJL|W: for now, make it just give you a few checkboxes for stuff like rp-pppoe and separate kernel module packages
12:40 < AllanonJL|W> ok, like ipw2200 etc?
12:40 < codeman> agaffney: you mean like the nforce and emu10k1 list in the handbook?
12:40 < agaffney> yeah
12:40 < AllanonJL|W> ok got it
12:41 < codeman> ok.  agaffney has the floor to tell us about partitioning
12:41 < AllanonJL|W> also, is there a way for me to have access to the gtkfe cvs directory other than web access or agaffney? hehe
12:42 < agaffney> uhh, partitioning works...somewhat :P
12:42 < codeman> what of it doesn't work?
12:42 < agaffney> everything appears to work, but there's only been very limited testing
12:42 < agaffney> mainly killing your Windows installs :)
12:43 < codeman> :)
12:43 < wolf31o2-work> how so?
12:43 < codeman> have you removed the ability to resize fat32 or are you just gonna give a huge warning?
12:43 < wolf31o2-work> because of the resize or?
12:43 < agaffney> wolf31o2-work: yes, fat32 + lba + resize = BADNESS
12:44 < agaffney> wolf31o2-work: not our issue or a parted issue but a fundamental FAT32 issue
12:44 < codeman> i saw some parted ppl in here, what were they helping with?
12:44 < agaffney> not parted
12:45 < agaffney> plors wrote gparted
12:45 < codeman> oh
12:45 < codeman> esammer: still not back?
12:46 < agaffney>                                                          _             _ 
12:46 < agaffney>   ___  ___  __ _ _ __ ___  _ __ ___   ___ _ __ _   _ __ (_)_ __   __ _| |
12:46 < agaffney>  / _ \/ __|/ _` | '_ ` _ \| '_ ` _ \ / _ \ '__(_) | '_ \| | '_ \ / _` | |
12:46 < agaffney> |  __/\__ \ (_| | | | | | | | | | | |  __/ |   _  | |_) | | | | | (_| |_|
12:46 < agaffney>  \___||___/\__,_|_| |_| |_|_| |_| |_|\___|_|  (_) | .__/|_|_| |_|\__, (_)
12:46 < agaffney>                                                   |_|            |___/   
12:46 < wolf31o2-work> agaffney: remove the resize capability?
12:46 < BenUrban> oh dear god
12:46 < samyron> LOL
12:46 < wolf31o2-work> by the way... how well does qtparted work? or is it the exact same problem?
12:46 < codeman> yeah agaffney you didn't answer my question b4
12:46 < agaffney> wolf31o2-work: I plan on disabling all resizing support for the first release
12:46 < samyron> agaffney: figlet rocks
12:47 < codeman> wolf31o2-work: as said above it's an issue with FAT32's design.
12:47 < agaffney> but it will still be able to deal with existing partitions
12:47 < wolf31o2-work> ok
12:47 < codeman> agaffney: and what filesystems DO you support?
12:47 < plors> agaffney: afaik a lot of fat32 resizing problems have been fixed in the lates libparted release (1.6.19)
12:47 < wolf31o2-work> so basically, it is BYOFS (Bring Your Own Free Space) or we wipe it all
12:48 < agaffney> codeman: anything that parted can detect and we'll have a tool for dealing with on the LiveCD
12:48 < agaffney> wolf31o2-work: you'll be able to delete and existing partition and create new partitions in its space
12:48 < agaffney> s/and/an/
12:48 < codeman> plors: when was that released?
12:49 < wolf31o2-work> right... that's pretty much what I said
12:49 < plors> codeman: 28 Nov
12:50 < agaffney> does npmccallum have gentoo-x86 commit access?
12:50 < codeman> agaffney: have you had a chance to check out the 1.6.19?
12:50 < agaffney> codeman: no, I didn't know it existed
12:50 < wolf31o2-work> agaffney: I do, if you guys need something
12:50 < agaffney> wolf31o2-work: I'm thinking about someone to take over maintainership of parted
12:50 < agaffney> since base-system basically told me to go fsck myself
12:50 < codeman> agaffney: well unless you see miracles fixed in that ver i suggest sticking with the current plan
12:51 < wolf31o2-work> agaffney: and it already looks like the livecd herd (which is just me, by the way) will be taking maintainership of the parted/pyparted stuff
12:51 < BenUrban> agaffney: pun intended?
12:51 < agaffney> BenUrban: no
12:51 < BenUrban> heh
12:52 < codeman> agaffney: is the partitioning data structure at least set in stone?
12:53 < codeman> like is the xml solid?
12:53 < agaffney> codeman: unless someone can give me a damn good reason to change it
12:53 < codeman> i think i gotta update the bootloader stuff w/ it since i was still using the fstab structure for that
12:53 < plors> codeman: i've tested fat32 quite thorough with 1.6.19 and it seems to be functional
12:54 < plors> no dataloss etc, no errors afterwards
12:54 < codeman> plors: with winxp?
12:54 < agaffney> plors: have you tested it on a drive >8.4GB?
12:54 < plors> dunno what ut does to bootsectors though..
12:54 < plors> agaffney: yup
12:54 < plors> with various cylindersizes as well
12:54 < agaffney> plors: the problem we were experiencing was related to LBA
12:54 < codeman> i hadn't lost any data, it just wouldn't boot the partition anymore
12:54 < wolf31o2-work> 2004-11-28  Andrew Clausen <clausen-mXXj517/[email protected]>
12:54 < wolf31o2-work> * libparted/fs_fat/boot_sector.c (fat_boot_sector_analyse): if the
12:54 < wolf31o2-work>   file system's CHS geometry is insane, then let the user fix it.
12:54 < wolf31o2-work>   This is important, because Linux refuses to mount FAT file systems
12:54 < wolf31o2-work>   with insane CHS geometry.  Moreover, a few recent versions of Parted
12:54 < wolf31o2-work>   wrote some insane values here, so this allows users to fix the damage.
12:55 < agaffney> plors: the resized fat32 partition still booted if the drive was <8.4GB
12:55 < plors> and didn't when > bigger ... ?
12:55 < agaffney> right
12:56 < plors> hmmz, cant test that over here :/
12:56 < agaffney> Windows XP died with a "Can't find NTLDR" error and no about of 'fixboot' and 'fixmbr' would fix it
12:56 < plors> otoh, after the install they have to set the mbr anyway
12:56 < agaffney> s/about/amount/
12:56 < plors> whit lilo/grub
12:56 < plors> so who cares what happens to the bootstuf..
12:56 < samyron> agaffney: i figured out how to fix that
12:56 < plors> oh, that's serious
12:56 < samyron> agaffney: well.. kinda
12:56 < wolf31o2-work> not really... both lilo and grub can be configured to chainload windows
12:57 < wolf31o2-work> but windows still has to be able to boot
12:57 < samyron> agaffney: I had to copy NTLDR and NTDETECT.COM to my OTHER FAT32 partition.. and then it booted
12:57 < agaffney> :-/
12:57 < plors> wolf31o2-work: you're right, as a nonwindows user i didn't realize that :)
12:57 < samyron> then you had to get into windows, go to the disk manager and set C: back to "System"
12:58 < samyron> but I think that was grub's problem
12:58 < agaffney> it appears that only 2K/XP were affected by this problem
12:58 < agaffney> 98SE still booted after a resize
12:58 < agaffney> in my tests
12:58 < codeman> 2k was ntfs
12:58 < agaffney> 2K and XP can both be installed on FAT32
12:59 < codeman> i know.  i meant my laptop 2k was ntfs
12:59 < codeman> and it resized fine
12:59 < agaffney> oh, ok
12:59 < agaffney> I was just referring to 2K in general
12:59 < codeman> i see that now
13:00 < codeman> ok, so since its 2:00 lets wrap this up
13:00 < codeman> FE looks like making great progress
13:00 < codeman> i will hopefully get the dialog one functional w/ some help
13:01 < codeman> and then we can work on fine-tuning the notification stuff
13:01 < agaffney> codeman: I'll be happy to help you with it, but my focus will be on the gtkfe
13:01 < codeman> i can't really say more about the notification system b/c i'm way behind in the code and haven't been able to catch up enuf
13:02 < codeman> but will we be able to have a progress bar of some sort?
13:02 < codeman> just based on the # of steps?
13:02 < samyron> sure.. won't be meaningful
13:02 < samyron> but we can
13:02 < agaffney> yeah, we'll just have to add a way to get the # of install steps from GLIClientController
13:02 < codeman> length()?
13:03 < agaffney> the fe shouldn't have direct access to that list
13:03 < agaffney> it will be trivial to add an accessor function
13:03 < codeman> agaffney: wait the # of install steps is coming from the individual arch template
13:04 < agaffney> codeman: yes, but the fe doesn't access the ArchTemplate directly
13:04 < agaffney> the fe accesses the ArchTemplate via GLIClientController
13:04 < codeman> so where does the accessor go?  clientcontroller or archtemplate?
13:04 < esammer> sorry - i'm back. got held up.
13:04 < agaffney> esammer: well not we're almost done :)
13:04 < agaffney> *now
13:04 < codeman> esammer: oh man you better read fast :)
13:04 < BenUrban> heh
13:05 < codeman> agaffney: just reminding you that you said you had somewhere to be around now.... 
13:06 < agaffney> yeah, I can wait a little while before I leave
13:06 < codeman> k, i dun wanna make you late for something
13:06 < agaffney> nah
13:06 < codeman> does the dialog successfully load profiles if you feed it one?
13:06 < agaffney> yes
13:07 < codeman> so i could load working samples and test the installer out theoretically
13:07 < agaffney> yes
13:07 < codeman> if we removed it from "pretend" mode?
13:07 < agaffney> yes
13:07 < agaffney> you may have to do some tweaking, but it should work
13:07 < codeman> alright i guess thats how i'll be testing from now on
13:08 < agaffney> remember, I've also got that install.py script
13:08 < codeman> yes i like that one :)
13:08 < codeman> ok, opening it up to general Q/A
13:09 < agaffney> what is the meaning of life?
13:09 < BenUrban> 42
13:09  * agaffney knew someone would say that :P
13:09  * spy|work looks for the infinity symbol
13:09 < codeman> that question was on my AI final
13:09 < codeman> i'm not joking
13:09 < BenUrban> lo
13:09 < BenUrban> l
13:09 < BenUrban> did you get it right?
13:10 < codeman> yes
13:10 < samyron> what'd you answer?
13:10 < samyron> "eat. sex. die"
13:10 < codeman> even though i have no idea where 42 comes from
13:10 < esammer> hhgttg
13:10 < BenUrban> the hitchhiker's guide to the galaxy
13:10 < BenUrban> wow, i type fast
13:10 -!- chief1983 [chief1983-39o3UGSKP6Pnn75J4d9gAMsRUnm6cuKJh13vi7wywA4@public.gmane.org] has joined #gentoo-installer
13:10  * z1nOnly|Work remindes everyone that z2nOnly wants to help, but not being a programmer, would like to offer services to test anything
13:10 < codeman> ah i should probably find that and watch it
13:10 < esammer> codeman: s/watch/read/g
13:10 < BenUrban> watch? try read
13:10 < esammer> ;)
13:11 < codeman> you mean you're two separate people?
13:11  * samyron has not yet read it either
13:12 < agaffney> everyone should read HHGTTG
13:12 < codeman> z?nOnly: got a box you can f- up in the name of scienc... gentoo?
13:12  * BenUrban agrees with agaffney 
13:12 < z1nOnly|Work> codeman: many
13:13 < codeman> kewl
13:13 < codeman> i'll be sure to throw you some scripts then
13:13  * agaffney needs to do some work on the backend partitioning code
13:13 < z1nOnly|Work> codeman: several laptops, one hardware raid box, and several server blade class machines
13:13 < agaffney> it can't handle MB and % yet
13:14 < codeman> agaffney: i thought that was the FEs job to convert
13:14 < plors> agaffney: feel free to ping me anytime about the parted stuff :)
13:14 < agaffney> codeman: no
13:14 < agaffney> codeman: we want to give people the ability to write profiles by hand
13:14 < agaffney> plors: oh, I will :P
13:15 < samyron> agaffney: size_in_mb * 1024 * 1024 = bytes
13:15 < samyron> all fixed
13:15 < samyron> :-)
13:15  * BenUrban wonders if plores meant that literally
13:15 < BenUrban> *plors
13:15 < agaffney> samyron: my partition code works in cylinders
13:15  * plors smacks BenUrban ;)
13:16  * BenUrban pings plors
13:16 < samyron> ah
13:16 < samyron> nm then
13:16 < codeman> ok anybody got anything else to share?
13:16 < plors> BenUrban: is there a meaning to the word 'ping' i dont know about ?
13:16 < BenUrban> not that i know of...
13:16 < plors> english isn't my first lang so it is possible :P
13:16 < plors> oh ok, nevermind :)
13:16  * agaffney doesn't get it
13:17  * codeman doesnt either.
13:17 < codeman> whatever
13:17  * BenUrban doesn't wither
13:17 < BenUrban> *either
13:17  * BenUrban can't spell today...
13:17 < codeman> esammer: any pearls of wisdom?  i know you've been wicked busy.
13:17 < agaffney> codeman: who is this esammer person you're talking about? :P
13:18  * BenUrban 's ping timed out :(
13:18 < codeman> i've been busy as hell too, but at least i stay in the channel to offer jokes every once and a while :)
13:19 < codeman> so these two drums and a cymbol fall off a cliff
13:19 < BenUrban> cymbal
13:19 < codeman> ... ba-dum ching.
13:19 < samyron> profile autoselection now works..
13:19 < esammer> codeman: i'm going to have to reread the scrollback. i think that you guys have been making great progress and more than anything, i'd like to just try to get back into the swing of things. at this point, enough people's eyes are going over things for them to be somewhat sane and do-able. we'll find what works and what doesn't and just go with things from there.
13:19 < samyron> well, based on whats in the client controller's config
13:19 < samyron> in case anyone cares
13:19  * agaffney cares
13:19 < samyron> agaffney: but you knew that already :-)
13:20 < agaffney> that doesn't mean I don't care
13:20 < agaffney> :)
13:20 < samyron> hehe
13:20 < samyron> true
13:20 < codeman> alright i guess that concludes the meeting