**** BEGIN LOGGING AT Sun Jul 24 02:59:56 2011
Jul 24 03:35:45 <dtzWill> Puffy_tp: we gotta get Mendeley up and running in time for me to use it in my quals prep! :)
Jul 24 05:20:39 <PuffTheMagic> dtzWill: im working on it as much as I can
Jul 24 05:38:47 <dtzWill> haha :D sokay i'm not taking quals until next spring, so we have some time. after fighting through my backlog of other projects, i do hope that i can help some as well :)
Jul 24 07:40:56 <dtzWill> rwhitby-pre/others: thoughts on how to make packages play nice across different devices?
Jul 24 07:41:13 <dtzWill> in the case of X server, ideally the scripts would know what platform they're on
Jul 24 07:41:19 <dtzWill> (or postinst or w/e sets things up properly)
Jul 24 07:42:08 <dtzWill> for example i want to push updated touhcpad keymappings, but can't since it requires a touchpad-specific Xsdl, inclusion of the keymappings themselves (nbd to ship everywhere), and modfying the set_keymap_later.sh script to use the correct keymaping
Jul 24 07:43:05 <dtzWill> one thought is have postinst or something just copy the correct Xsdl into the 'canonical' place (so scripts don't have to know), and ditto for keymappping
Jul 24 07:43:26 <dtzWill> I think rwhitby-pre mentioned that was his preferred solutoin to a similar problem earlier...
Jul 24 09:09:37 <cryptk> PuffTheMagic, have you put any thought of porting over evince and poppler >0.15
Jul 24 09:09:53 <cryptk> supposedly evince can do PDF annotations if you have poppler v0.15 or higher
Jul 24 09:10:06 <cryptk> and then we would have a full document viewer... not just PDF's...
Jul 24 09:10:32 <dtzWill> okular! \o/
Jul 24 09:11:22 * dtzWill wonders if okular could run using the qt stuff ported earlier
Jul 24 09:11:50 <dtzWill> i know qt is shipped with device now, but i'm unclear how it's used--for launcher and such, right? (not like the card stuff from earlier)
Jul 24 09:14:50 <cryptk> I wonder what the minimum libraries required are for evince and okular
Jul 24 09:16:46 * dtzWill nods
Jul 24 09:17:00 <cryptk> evence states that it requires "GNOME Platform libraries" but doesn't mention which ones
Jul 24 09:17:07 <cryptk> it can't possibly require them all
Jul 24 09:18:53 <cryptk> dtzWill, do you know if we can do file type associations with app catalog apps?
Jul 24 09:19:37 <dtzWill> i don't, but i'm the wrong guy to ask :)
Jul 24 09:27:53 <cryptk> evince definitely is going to require gtk
Jul 24 09:28:08 <cryptk> xpdf would be easier to port, but doesn't do annotations
Jul 24 09:34:35 <cryptk> it looks like xpdf will show annotations... I am unsure if it will let you edit or add them though
Jul 24 09:47:06 <cryptk> dtzWill, if I build Xserver from git will I be able to do capitol letters yet?
Jul 24 09:53:06 <dtzWill> cryptk: kinda, not really.
Jul 24 09:53:22 <dtzWill> package creation is borked in widk due to the upstart changes to make it work on 1.4.5
Jul 24 09:53:56 <cryptk> ahh
Jul 24 09:54:02 <cryptk> thta's fixable though right?
Jul 24 09:54:08 <dtzWill> probably, yeah.
Jul 24 09:54:14 <dtzWill> problem is i have it built with palm-package
Jul 24 09:54:16 <cryptk> just needs the package part updated
Jul 24 09:54:27 <dtzWill> and upstart and whatnot requires building the package ourselves
Jul 24 09:54:36 <cryptk> btw, I am checking out the source of jmonkeyengine
Jul 24 09:54:41 <cryptk> that has lwjgl in it...
Jul 24 09:54:51 <cryptk> guess what I'm gonna work on over the weekend...
Jul 24 09:55:02 <dtzWill> (I don't consider build.git part of the dev/test cycle, since it requires tagging the release i think? Maybe we can get special chase-head targets for testing.... idk)
Jul 24 09:56:10 <dtzWill> anyway if you build latest xserver, you'll have a touchpad-only Xsdl
Jul 24 09:56:15 <dtzWill> when combined with git of xkeyboard-config
Jul 24 09:56:25 <dtzWill> and modify set_keymap_later.sh to 'touchpad'
Jul 24 09:56:28 <dtzWill> things work
Jul 24 09:56:52 <dtzWill> aka it's there, but the 'push this to users' pipeline got SNAFU'd and i haven't had the resolve or time to fix it yet :)
Jul 24 09:57:07 <cryptk> gotcha
Jul 24 09:58:44 <dtzWill> still in "make something cool mode", especially with the more recent framebuffer goodness find :)
Jul 24 09:58:47 <dtzWill> :3
Jul 24 09:58:58 <cryptk> what framebuffer goodness find?
Jul 24 09:59:13 <dtzWill> cryptk: much faster blitting
Jul 24 09:59:24 <dtzWill> still not hw-accel'd, but for basic usage is _much_ better
Jul 24 09:59:38 <dtzWill> misc mplayer benchmark went from ~11.5s down to ~3.5s
Jul 24 10:00:04 <dtzWill> (And that includes mpeg-2 software decoding overhead, so the speedup might be even better than that ratio)
Jul 24 10:00:14 <dtzWill> informally, it's just much much better :)
Jul 24 10:00:35 <cryptk> nice
Jul 24 10:00:51 <dtzWill> however it's kinda hacky, and i'll (we'll) need to handle a couple edge cases to avoid small graphical glitches.
Jul 24 10:01:54 <dtzWill> but yeah, i might try it in my emulators too--and there it's pretty glorious in terms of being able to draw quickly
Jul 24 10:09:27 <cryptk> ok, I have the lwjgl native library source
Jul 24 10:09:42 <cryptk> which was not in the jmonkeyengine code, only compiled versions were
Jul 24 10:09:46 <cryptk> but I have found it
Jul 24 10:10:59 <cryptk> only thing is since it is made for java... it is an ant style build.xml that assembles a gcc command and runs it
Jul 24 10:11:18 <cryptk> but I should be able to rip out the important pieces and make a gcc command to use in widk
Jul 24 10:11:50 <cryptk> -lX11 -lXext -lXcursor -lXrandr -lXxf86vm
Jul 24 10:11:59 <cryptk> we don't have libXrandr do we?
Jul 24 10:12:03 <cryptk> or do we?
Jul 24 10:20:09 <dtzWill> cryptk: we sure do ^.^
Jul 24 10:24:34 <cryptk> ooh... lemme see what I can do here then
Jul 24 10:25:45 <cryptk> oooh... ant is part of the WIDK... no need to convert for testing then
Jul 24 10:28:21 <cryptk> woohoo... downloading the toolchains for WIDK at 5MB/s
Jul 24 10:29:03 <cryptk> taking longer to extract it than it did to download it, lol
Jul 24 10:30:00 <cryptk> SDK is downloading at 9MB/s....
Jul 24 10:30:05 <cryptk> man I love my connection at work
Jul 24 10:31:46 <cryptk> hrm... I kinda want to set up a dev server
Jul 24 10:31:57 <cryptk> just have the WIDK on it... get everything built and staged
Jul 24 10:32:02 <cryptk> then we can all use it
Jul 24 10:32:48 <cryptk> make some kind of a script where in everyones home dir they get the built WIDK bind mounted in, with an AUFS overlay on it
Jul 24 10:32:59 <cryptk> so they can make their own local changes, but the baseline stays the same
Jul 24 10:33:10 <cryptk> dtzWill, do you think that would be useful^^^
Jul 24 10:51:39 <cryptk> wtf, I can't get minecraft to come up at all
Jul 24 10:51:44 <cryptk> I haven't even changed anything yet
Jul 24 10:51:54 <cryptk> before I would at least see the login screen
Jul 24 10:56:12 <cryptk> bah... stupid python.org has an IPv6 address now... and my DNS is resolving to it... but I don't have an IPv6 addy to talk to it
Jul 24 10:59:00 <cryptk> or... python.orgis down
Jul 24 11:37:34 <cryptk> dtzWill, so I have libx11, libxext and libxrandr all staged... I need libXcursor and libXxf86vm
Jul 24 11:37:44 <cryptk> those aren't being supplied by anything else by chance are they?
Jul 24 11:46:53 <cryptk> bah... I hate it when source code doesn't come with something telling you what it depends on
Jul 24 12:08:25 <cryptk> I think I have libXfixes working... that's needed for libXcursor
Jul 24 12:55:24 <cryptk> dtzWill, what's the policy on adding new libraries to WIDK?
Jul 24 12:55:34 <cryptk> add them to nonworking first and then get someone else to test?
Jul 24 13:11:51 <cryptk> ok, libXcursor builds clean... no dependency problems anymore
Jul 24 13:13:05 <cryptk> giving libXxf86vm a test run now off a fresh clobber of packages/x/
Jul 24 13:16:45 <cryptk> and if that works, then everything I added is getting thrown into a folder in nonworking for now until I get home and someone else tests them
Jul 24 13:17:00 <cryptk> and preferably until I get minecraft working, because that will confirm that it actually works
Jul 24 13:32:44 <cryptk> ok libXcursor libXfixes libXxf86vm xf86vidmodeproto all pushed into nonworking, they all build, but they are all untested
Jul 24 14:10:54 <cryptk> ok, back at home... time for more work on minecraft
Jul 24 14:17:31 <cryptk> just freshly reinstalled the ubuntu-chroot... I want this to be reproducible
Jul 24 14:59:06 <PuffTheMagic> dtzWill: poppler can fiew them i didnt think it could edit them
Jul 24 14:59:13 <PuffTheMagic> real annotations that is
Jul 24 14:59:22 <PuffTheMagic> it has always had its own external format
Jul 24 15:31:54 <cryptk> having to decipher an ant script and turn it into gcc
Jul 24 15:41:15 <PuffTheMagic> cryptk: im not sure if you remember my preyourmindsite... but i generated feeds on the fly for a half a dozen packages and having that feed included in preware by default destroyed the server
Jul 24 15:41:39 <PuffTheMagic> my serverthat was
Jul 24 15:53:01 <PuffTheMagic> dtzWill: we could prob just do a glib -> sdl thing
Jul 24 15:53:06 <cryptk> PuffTheMagic, this isn't going to be run on my personal server
Jul 24 15:53:09 <PuffTheMagic> for poppler
Jul 24 15:53:14 <cryptk> I am just developing it there
Jul 24 15:53:38 <cryptk> when it is all done, if rwhitby approves, it will go on the ipkg servers
Jul 24 15:54:04 <cryptk> I am going to use it to build a homebrew version of the online app cat as well
Jul 24 15:54:35 <cryptk> bah.. I don't know enough ant or make to get this darn thing to build
Jul 24 15:54:43 <PuffTheMagic> im just saying that if it were in preware it will prob bring down any server
Jul 24 15:55:04 <cryptk> really... because the ipkg servers handle their load just fine
Jul 24 15:55:04 <PuffTheMagic> s/in preware/use in preware/
Jul 24 15:55:30 <cryptk> and each request would be very small.... the thing about the API, it allows for incremental updates
Jul 24 15:55:34 <PuffTheMagic> ya but the content its dynamic and creted on the fly
Jul 24 15:55:58 <PuffTheMagic> i cant talk its early...
Jul 24 15:56:04 <PuffTheMagic> the current feeds are static i mean
Jul 24 15:56:11 <cryptk> I ran it on a 1GB RAM server at work last night and I was able to handle 500 concurrent connections
Jul 24 15:56:23 <cryptk> serving all requests in well under 1 second
Jul 24 15:56:43 <cryptk> the data is cached... the mysql tables are cached... the PHP opcode is cached
Jul 24 15:56:48 <cryptk> it runs very quickly
Jul 24 15:56:48 <PuffTheMagic> people sync preware 20 times a day
Jul 24 15:56:53 <PuffTheMagic> and how many users do we have?
Jul 24 15:57:31 <cryptk> http://munin.nas-admin.org/nas-admin.org/milla/nginx_request.html
Jul 24 15:57:33 <cryptk> look at that
Jul 24 15:58:01 <cryptk> milla (the main ipkg server) sits at below 10 requests/second typically
Jul 24 15:58:39 <cryptk> and those requests are serving large Packages.gz files... several of them, every time someone refreshes their feeds
Jul 24 15:59:25 <cryptk> I was handling 500 concurrent connections with a response time of about 1 second... which means 500 requests per second
Jul 24 15:59:31 <cryptk> FAR FAR FAR above what milla sees
Jul 24 16:00:01 <cryptk> and the packages themselves are not going to be served from the API server... it just offloads the process of parsing the Packages.gz files
Jul 24 16:00:56 <cryptk> your average request to the API server will have only a few packages returned... queried from a database that is only a couple of MB... and well indexed... and cached
Jul 24 16:01:12 <cryptk> I think you are overestimating the amount of processing power required to do this
Jul 24 16:02:09 <cryptk> you were generating feeds on the fly, my data is already in place, it isn't generating it on the fly... it just pulls it from a database and formats it into JSON
Jul 24 16:10:47 <cryptk> even the varnish stats on milla show an average request rate of ~3/s with a spike up to ~6.5 and a spike up to 8
Jul 24 16:10:52 <cryptk> still far below what I tested at
Jul 24 20:26:28 <PuffTheMagic> dtzWill, well i just made some progress with the database
Jul 24 20:27:34 <PuffTheMagic> when the document details are fetched it does a query on the db by type and id and tries to merge the new data with the return of the query
Jul 24 20:27:47 <PuffTheMagic> if it fails then it pushes a new item into the db
Jul 24 20:28:00 <PuffTheMagic> this way meta data can be updated with out creating a new entry
Jul 24 20:28:11 <PuffTheMagic> as long as the document id doesnt change
Jul 24 20:29:44 <PuffTheMagic> you can also do queries with change notifications
Jul 24 20:29:58 <PuffTheMagic> so we can do a seach that returns all documents
Jul 24 20:30:08 <PuffTheMagic> and then when new items are added it will return them
Jul 24 20:30:37 <PuffTheMagic> but i think there is a list widget that is driven by database too
Jul 24 20:38:58 <PuffTheMagic> no we can get sorting and searching for free
Jul 24 21:34:41 <rwhitby> morning
Jul 24 21:34:52 <ka6sox> morning rwhitby
Jul 24 21:37:21 * rwhitby is in the air between LAX and PDX
Jul 24 21:39:07 <halfhalo> nice.
Jul 24 21:39:43 <PuffTheMagic> dtzWill, yup, this DbList is awesome
Jul 24 21:39:59 <PuffTheMagic> got it sorting by year with a toggle for ascending or descending
Jul 24 21:40:14 <PuffTheMagic> will add sorting by author and publication_source too
Jul 24 21:47:20 <rwhitby> ka6sox: pushing tailor now
Jul 24 21:47:38 <rwhitby> Tailor pushed.
Jul 24 21:48:04 <rwhitby> "make device test" in /srv/applications/tailor to test it
Jul 24 21:48:48 <rwhitby> ka6sox: are you able to test it?
Jul 24 22:01:51 <ka6sox> i will in a little while...taking care of my son right now.
Jul 24 22:03:09 <rwhitby> no worries.
Jul 24 22:03:23 <rwhitby> it doesn't actually resize anything yet, I'm creeping up on that.
Jul 24 22:04:08 <rwhitby> If I had some more elbow room on this flight I'd continue working on it ...
Jul 24 22:31:28 <ka6sox> let me pull down what you have so far and add the changes I've made
Jul 24 22:33:38 <ka6sox> at some point we will need to either decide that a value of Zero MB for the ext3fs triggers the question" do you want to remove the ext3fs" or an explicit button for that.
Jul 24 22:39:29 <ka6sox> uh....all the scripts I wrote are gone now...
Jul 24 22:39:40 <ka6sox> I guess I could recreate them.
Jul 24 22:39:47 <rwhitby> I haven't removed anything
Jul 24 22:39:57 <ka6sox> when i did a pull they all disappeared
Jul 24 22:39:57 <rwhitby> I didn't see any new stuff pushed by you
Jul 24 22:40:17 <ka6sox> I mean I can't find them period.
Jul 24 22:40:33 <rwhitby> were they checked in?
Jul 24 22:40:46 <ka6sox> yes, I pushed them 3days ago
Jul 24 22:41:04 <ka6sox> I can rewrite them...
Jul 24 22:41:07 <rwhitby> show me the commit on the git repo web site?
Jul 24 22:41:18 <ka6sox> jas....
Jul 24 22:41:57 <rwhitby> I see 4 scripts in ./sbin/
Jul 24 22:42:08 <rwhitby> from http://git.webos-internals.org/applications/tailor/commit/?id=08cea580882f2d6ac74ec65a053c4106f5b3bed8
Jul 24 22:43:09 <ka6sox> wierd...pulled again...and now they are tehre.
Jul 24 22:43:27 <ka6sox> before ls sbin/ showed me what was in bin/
Jul 24 22:43:50 <ka6sox> okay now I can work on pushing up my changes.
Jul 24 22:53:00 <ka6sox> rwhitby, which way should we deal with that?
Jul 24 22:54:57 <ka6sox> laters
Jul 25 00:45:23 <PuffTheMagic> dtzWill, you can now sort your library by publication outlet :D
Jul 25 00:47:20 <PuffTheMagic> dtzWill, http://ompldr.org/vOW04MA
Jul 25 01:27:52 <dtzWill> PuffTheMagic: nice!! :D ty
Jul 25 01:28:07 * dtzWill will look at it and non-hilighted backlog in a bit
Jul 25 01:34:15 <dtzWill> oh yay db goodness
Jul 25 01:36:46 <rwhitby> dtzWill: did you get a chance to look at that PWD thing?
Jul 25 01:37:02 <rwhitby> (cause I'm just about to try a make stage again)
Jul 25 01:39:18 <dtzWill> rwhitby: should be fixed, yes.
Jul 25 01:39:27 <dtzWill> bbiab, lmk if you have any issues
Jul 25 01:39:28 <dtzWill> o/
Jul 25 01:39:43 <rwhitby> dtzWill: did you push the fix?
Jul 25 01:40:17 <rwhitby> cp: cannot stat `../xkeyboardconfig/build/armv7/keymap/touchpad': No such file or directory
Jul 25 01:40:33 <rwhitby> from a make stage in xserver-package
Jul 25 01:40:40 <dtzWill> sigh, i dind't. very sorry.
Jul 25 01:41:16 <rwhitby> no worries
Jul 25 01:41:44 <dtzWill> oh that's a different change, i guess i didn't tag it.
Jul 25 01:42:05 <dtzWill> sorry the upstart stuff broke the widk packaging stuff so i kinda dropped the ball on the other things
Jul 25 01:42:34 <dtzWill> (and i can'tto fix it+crossdevices issues+the rest atm)
Jul 25 01:43:02 <dtzWill> hmm mabye i can tag that
Jul 25 01:43:06 <dtzWill> doesn't _hurt_ since it's unused
Jul 25 01:45:29 <dtzWill> rwhitby: pushed
Jul 25 01:45:43 <dtzWill> slightly untested since i am lacking disk space for one reason or the other. sorry about the trouble.
Jul 25 01:48:20 <ka6sox> rwhitby, what toolchain does tailor use?
Jul 25 01:48:38 <rwhitby> normal
Jul 25 01:50:00 <ka6sox> not widk one?
Jul 25 01:50:41 <rwhitby> it uses palm SDK on osx and linux
Jul 25 01:51:53 <rwhitby> same as all the other Makefiles in /srv/applications/*
Jul 25 01:53:48 <ka6sox> kk
Jul 25 02:16:06 <jacques> argh
Jul 25 02:16:15 <jacques> I thought they had fixed the screenshot issue
Jul 25 02:16:42 <jacques> I'm trying to take screenshots, but they don't look right - they don't look like the screen
Jul 25 02:22:55 <jacques> I bet it's because it's 16-bit with depth buffer
Jul 25 02:58:27 <linuxjacques> dtzWill, about?
**** ENDING LOGGING AT Mon Jul 25 02:59:57 2011