**** BEGIN LOGGING AT Mon Sep 19 02:59:57 2011 Sep 19 03:02:44 yawn Sep 19 03:05:09 * dtzWill passes ka6sox-away some coffee Sep 19 03:05:22 need some Sep 19 03:05:35 but I guess it will be DDP Sep 19 03:46:24 scoutcamper, so Xapps is now xapps Sep 19 03:46:52 ok Sep 19 03:46:52 so you need to make sure that everything is either already pushed up there or we can get it to the new one. Sep 19 03:47:03 then I will clone and make some mode Sep 19 03:47:08 s/mode/mods Sep 19 03:47:59 u want the shimm there? Sep 19 03:48:40 shim will stay where it is Sep 19 03:48:52 we are making apps chroot specific Sep 19 03:49:12 ok Sep 19 03:49:16 look @ the Icons page in the wiki Sep 19 03:49:19 what do you want in the repo? Sep 19 03:49:43 look for the ones that have the ubuntu logo lower left and X in the lower right corner Sep 19 03:50:56 ok Sep 19 03:51:49 and start with Xchat as our first victim Sep 19 03:52:03 then I will finish up the stuff necessary Sep 19 03:52:21 so the patches we did for the ubuntu chroot stand Sep 19 03:52:31 and we get rwhitby to include them. now Sep 19 03:53:13 those include the additional user + adding the shim. Sep 19 03:53:58 ka6sox-away, http://git.webos-internals.org/x11/xapps/ Sep 19 03:54:13 okay let me put the things I know need to be there in there. Sep 19 03:54:22 for Xchat Sep 19 03:55:00 there is a small script that goes in /root/scripts so we can setup the user rights to things can get saved. Sep 19 03:55:09 BTW, dtzWill and I are thinking of adding service methods to xecutah to detect if an X server is running on a port, and a separate method to start one on a specific port. Sep 19 03:55:53 okay, I was hoping we could make it so that xecutah would start a xserver on a port and assign it to the next app that needs one. Sep 19 03:55:57 so that XTerm would change from being a changelog-only app into an app that starts an xserver and runs an xterm and then exits, allowing you to start a second xterm by clicking on the same app again Sep 19 03:56:17 okay thats what I was hoping Sep 19 03:56:18 good. Sep 19 03:56:26 no timeframe though :) Sep 19 03:56:30 kk Sep 19 03:56:41 for now its hardcoded but thats an easy mod Sep 19 03:56:56 changes the display= var Sep 19 03:57:08 yeah i'd say the two works are pretty easily combinable Sep 19 03:57:24 or at least the xapps stuff should be able to leverage those extensions to xecutah in a straightforward way Sep 19 03:57:34 good. Sep 19 03:57:46 right Sep 19 03:57:47 those changes would make it so that we can run a number of xapps Sep 19 03:58:03 including xterm. Sep 19 03:59:57 okay, so for xapps we have converted so that it will have a path that conforms to /media/ext3fs/wherever_the_chroot_is Sep 19 04:00:26 for the existing natty one thats fixed by the one rwhitby created as a package Sep 19 04:00:48 rwhitby: so i looked and there's a 'xwininfo' in widk already Sep 19 04:00:55 that should give us what we want Sep 19 04:01:10 modulo LD_LIBRARY_FLAGS telling it where to find the libraries it wants Sep 19 04:01:18 dtzWill: oh, could that do server and window detection ? Sep 19 04:01:19 xwininfo -root -display $DISPLAY Sep 19 04:01:37 oh connect, return value is good else it's bad Sep 19 04:01:39 dtzWill: we should just include it with xserver Sep 19 04:01:52 it does other things too, but it also does what we want and is small Sep 19 04:02:05 well it's 128K lol whoops Sep 19 04:02:09 (oh well) Sep 19 04:02:24 mkay i can package it up Sep 19 04:02:30 it comes and goes though right? Sep 19 04:02:50 what comes and goes? Sep 19 04:02:57 that xwininfo Sep 19 04:03:16 what do you mean comes and goes? Sep 19 04:03:20 nm Sep 19 04:03:22 continue Sep 19 04:03:35 lol okay, i just don't follow Sep 19 04:03:48 doesn't matter Sep 19 04:03:53 lol okay Sep 19 04:03:53 :) Sep 19 04:03:56 * dtzWill lets it go Sep 19 04:04:38 so I will put the startup script in /root/scripts and the chroot call will be in the wrapper. Sep 19 04:05:25 this reminds me i need to enable those alpha feeds Sep 19 04:05:38 so i can start pushing broken shit to them }:D Sep 19 04:05:52 rwhitby, when we call the upstart script can it accept an argument that is the display=:X info? Sep 19 04:06:05 becuase right now its hardcoded. Sep 19 04:06:37 or will the wrapper script need to find the correct xdisplay? Sep 19 04:07:08 not sure how to pass parameters to upstart scripts Sep 19 04:08:15 are we still trying to be 1.4.5 compatible? Sep 19 04:08:57 yep, xecutah has many 1.4.5 users Sep 19 04:09:17 so only 1 script call then. Sep 19 04:09:34 how does 1.4.5 affect things? Sep 19 04:09:39 since it can't do 2 iirc. Sep 19 04:10:54 not sure what you're referring to ... Sep 19 04:11:31 what was the issue with trying to do 2 things with executah for 1.4.5 clients? or is that gone? Sep 19 04:12:19 oh, I think that was the xserver startup script not returning Sep 19 04:12:28 I think dtzWill has done some magic to fix that now Sep 19 04:12:33 okay Sep 19 04:12:59 okay then non-issue. Sep 19 04:13:08 well, should be tested. Sep 19 04:13:30 kk Sep 19 04:13:52 then I will finish up my bits to make xchat be ready for the catalog Sep 19 04:13:58 lol do you really ahve to reinstall preware? :) Sep 19 04:14:05 ? Sep 19 04:14:12 for testing feed enabling Sep 19 04:14:17 dtzWill: yep, feeds are created in the postinst Sep 19 04:14:31 mkay Sep 19 04:14:38 guess that's easier than adding the feeds directly myself Sep 19 04:14:38 xD Sep 19 04:16:29 * dtzWill runs postinst directly Sep 19 04:16:39 * ka6sox-away too Sep 19 04:17:20 curiously that got them added but not enabled, fun Sep 19 04:17:27 oh Sep 19 04:17:34 * dtzWill reads rest of wiki where that's what it's supposed to do Sep 19 04:17:35 also, nice :) Sep 19 04:17:40 heh, someone didn't read the whole page before starting ;) Sep 19 04:17:44 ;) Sep 19 04:17:46 not at all Sep 19 04:17:47 xD Sep 19 04:17:54 "I'll figure it out if it breaks" Sep 19 04:17:56 :[ Sep 19 04:18:03 <--- not agood testing candidate Sep 19 04:18:23 lucky you're the developer, or we'd kick you out ;) Sep 19 04:18:24 oh wonderful, we've split the feeds into {alpha,beta}x{apps,kernels,optware,patches} Sep 19 04:18:25 exceellent Sep 19 04:18:39 so i can have alpha apps but not alpha kernels, woo Sep 19 04:18:51 * ka6sox-away likes that too. Sep 19 04:19:25 * ka6sox-away wonders if he would be a candidate for being kicked out too...not good Dev. Sep 19 04:20:41 as soon as this SCALE meeting is over I'll finish up xchat for ubuntu and publish it to the alpha feed :D Sep 19 06:18:20 rwhitby: when to bump the preware version number? each new feature? Sep 19 06:18:37 * Xuzz noticed the in-app changelog is slightly different from the one in the package description Sep 19 06:18:54 Xuzz: oh, the last two entries are intentionally swapped. Sep 19 06:19:10 but should be the same in both places Sep 19 06:19:44 Xuzz: where do they differ? I see the same here. Sep 19 06:20:22 um, that's a good question. Sep 19 06:20:40 oh, i think it was the changelog for my last two changes was slightly different or something Sep 19 06:21:03 so an Alpha would be something like 0.0.1 for a version #? Sep 19 06:29:48 Xuzz: 1.7.5 and 1.7.6 look the same (modulo formatting) Sep 19 06:29:56 ka6sox-away: doesn't matter Sep 19 06:29:56 okay, nevermind then Sep 19 06:35:08 does anyone else generally find the keyboard use a bit..."thick" on X? Sep 19 06:35:19 was trying to do some semi-serious programming the other day, and found it more than a little difficult Sep 19 06:35:30 and i doin't think it's just the keyboard.... anyway, anyone else have similar experience? Sep 19 06:35:35 seems to miss some keys and the like Sep 19 06:35:40 you mean the vkb or BT one? Sep 19 06:35:43 i mean bt Sep 19 06:35:51 i'm not crazy enough to try to program with the vkb Sep 19 06:35:52 haha Sep 19 06:36:03 its fine...except for the lack of caps working consistently. Sep 19 06:36:17 and I still can't get arrows or ctrl to work. Sep 19 06:38:23 that's a bummer. maybe your keyboard does things differently. you're definitely running the latest? should've fixed that quite some time ago Sep 19 06:38:24 bummer :( Sep 19 06:42:51 let me poke around again Sep 19 06:42:59 but last time I checked it didn't Sep 19 06:43:09 bleh...3hr meetings online Sep 19 06:44:22 Socket_CAN is SET. Sep 19 06:45:30 whoops...thats for the panda_car_puter Sep 19 07:41:30 dtzWill, yeah i find myself needing to press buttons on the BT keyboard "all the way" to register some keys Sep 19 10:32:55 does anyone know is it possible to implement queue sort of thing in JS services? i.e. so that service calls are processed one by one and not by starting a new instance for every service call? Sep 19 10:34:31 I have a problem with MS when there are multiple calls for the service via activities in short time of period, the service is started twice and the first call does not have time to finish (write changed data into db) before the second call is processed (which uses outdated data from db) Sep 19 10:34:43 I would like to find a way to fix that Sep 19 10:40:12 sconix: maybe a global future, and service calls just add onto that global future? Sep 19 10:42:17 how can you make a global future? Sep 19 10:42:30 that most likely would work perfectly Sep 19 10:45:13 will have to see if I can do such a thing, it really would be what I am after Sep 19 11:07:50 sconix: perhaps make it in the service assistant for the node service, rather than in the command assistant? Sep 19 11:08:12 dtzWill: ping Sep 19 11:08:26 rwhitby: thanks I will look into that Sep 19 11:08:46 rwhitby: pong-ish Sep 19 11:09:00 hmm - I have upstart scripts in /var/palm/event.d/ but they are not being recognised by upstart Sep 19 11:09:27 hmm, like the x users are reporting. Sep 19 11:09:30 updated 4 packages in Preware. all of the packages had upstart scripts. all scripts got installed correctly. 2 are recognised, 2 are not. Sep 19 11:09:39 o_O Sep 19 11:09:54 dtzWill: yes, xserver and xterm are the two that are not recognised Sep 19 11:10:38 dtzWill: BTW, I've added multi-display support to xserver, xterm, ubuntu, debian Sep 19 11:11:24 all the various scripts should correctly pass through a DISPLAY environment variable instead of forcing :0.0 (they all still default to :0.0 if DISPLAY is not set) Sep 19 11:12:00 xecutah will pass display to the service, which will set it as an environment variable when it calls initctl start ... Sep 19 11:12:02 great :D Sep 19 11:12:23 http://netsplit.com/2010/12/03/event-matching-in-upstart/ is where I got the idea Sep 19 11:12:50 sorry i meant to add at least xwininfo, but am sucked into this paper deadline atm. will help push forward on xecutah detecting which displays have x servers in a bit :) Sep 19 11:12:50 to start with, I will make xecutah increment the display number each time you start XServer Sep 19 11:12:52 * dtzWill looks at link Sep 19 11:13:11 yep, you can just do that as a cheap allocation policy at first Sep 19 11:13:13 the method I have devised does not need the status call Sep 19 11:14:24 but now I'm in danger of being distracted by this upstart script recognition problem Sep 19 11:14:40 at least now I know the symptom, but not yet the root cause. Sep 19 11:14:56 well if rebooting fixes it for everyone maybe that's best >_> Sep 19 11:14:56 lol Sep 19 11:15:28 yes, rebooting fixes it. but that's a crude solution for a Linux box. Sep 19 11:15:37 heh yeah :// Sep 19 11:16:46 hmm, reboot didn't fix it. Sep 19 11:17:34 rwhitby: are you hoping to have well-crafted upstart jobs that run for each x server and their status being the corresponding server's status? Sep 19 11:18:23 dunno Sep 19 11:19:24 i suppose that'd work fine, although i feel like it shouldn't. anyway, until either of us (or someone else) sees why that's the case it probably would work fine Sep 19 11:20:18 (this seems to be functionally equivalent to simply tracking what processes we've started, even by grepping 'ps' results, if quite a bit cleaner, which is insufficient for general X usage appraently judging by how paranoid their startup stuff is) Sep 19 11:20:32 but i'm willing to be a good deal of their paranoia is for things we can't more or less safely dismiss Sep 19 11:20:53 for example, if a user hand-starts an x-server that we don't know about it, i'm not sure we should go toooo far out of our way to handle that super well Sep 19 11:25:31 hmm reboot doesn't fix it?? bah Sep 19 11:25:33 damn. looks like the environment stuff is only in upstart 0.5.0 - we have upstart 0.3.8 Sep 19 11:34:09 aww, damn indeed. Sep 19 11:49:50 looks like initctl emit can send an event with env variables Sep 19 11:50:26 gad damn how have I missed the fact that service can have an assistant as well :D Sep 19 11:50:41 sconix: it's well hidden Sep 19 11:50:44 this looks just what I need to make MS work 100% even with many trigers Sep 19 11:51:41 rwhitby: jeah I have tried to look into many services to see if they have a such thing, once I found the code part to access service assistant but it did not work, which now makes sense since I did not address the json file side Sep 19 11:52:10 well better late than never, at least now I can finally fix the number one issue in the MS service to make it really realiable Sep 19 12:06:19 dtzWill: so I can emit an upstart event which has associated environment variables, and can test for those variables in the start on stanza, but cannot seem to get them passed down to the script that is executed :( Sep 19 12:07:48 have an upstart job for each server/match? and hardcode the display? :3 Sep 19 12:08:47 dtzWill: yeah, already thought of that. unfortunately would need to be replicated for xterm, ubuntu, debian, etc ... Sep 19 12:09:28 xecutah-demux-generator Sep 19 12:09:28 lol Sep 19 12:13:59 dtzWill: however, if xecutah.sh is reliable, then perhaps we can cut upstart out of the picture now Sep 19 12:15:10 i thought upstart was required for 1.4.5 devices? Sep 19 12:15:30 also, i have not tested on 1.4.5 myself although it should work. guess we'd have heard if i broke it, that update was pushed a while ago O:) Sep 19 12:15:30 well, that was because xecutah.sh was not returning Sep 19 12:15:35 interesting. Sep 19 12:15:50 the forker thing was part of the problem, right? Sep 19 12:15:56 and now it's been removed Sep 19 12:17:18 sounds good to me Sep 19 12:17:30 dodges the upstart job not being registered issue too xD Sep 19 12:18:10 however, using upstart start on .... events, and individual scripts for each display, perhaps we can get xterm to start the server automatically Sep 19 12:19:56 dtzWill: so, how do we reduce the startup and keymapping etc time for xserver? Sep 19 12:20:20 well startup time atm i think is dominated by hitting tweaks i think Sep 19 12:20:25 currently it takes 5 seconds for initctl to return Sep 19 12:20:35 else it's the minimum xserver startup plus the xkb stuff Sep 19 12:20:53 i think around 1 or 2 seconds are the tweaks dbus calls Sep 19 12:21:04 then the xserver is started Sep 19 12:21:23 and 16 seconds if there is an error :) Sep 19 12:21:25 as soon as it can accept connections, it sends signal to the xecutah.sh which runs the xkb stuff Sep 19 12:21:39 and yes 15 seconds is the timeout in xecutah.sh Sep 19 12:22:13 i wanted it to not be unBEARably long but long enough that a completely cold boot on a loaded pixi wouldn't be considered failure Sep 19 12:22:31 rwhitby: subsequent launches are faster, right? Sep 19 12:26:34 dtzWill: launches of what? Sep 19 12:26:51 oh of xserver Sep 19 12:27:03 like if it's 5 seconds first time, if you close it and try again is it stil a full 5 seconds? Sep 19 12:27:18 I'd expect a good part of that is pulling in the various libraries that should be happily cached on later loads Sep 19 12:27:56 3.8 seconds Sep 19 12:28:32 heh Sep 19 12:28:33 kk Sep 19 12:28:43 didn't mean bust out your stopwatch xD, but that works. Sep 19 12:28:49 okay so still a decent amount of time Sep 19 12:28:53 time initctl :) Sep 19 12:29:20 but not too bad that it can't stay that way for the moment Sep 19 12:30:30 dtzWill: can we make the xserver xecutah.sh check for a running server on that DISPLAY and exit immediately with success if there is one already running? Sep 19 12:30:47 rwhitby: i can make it do so, yes. Sep 19 12:30:58 if you wanna delegate that so that something you're buliding works i'm cool with that Sep 19 12:31:27 then the xterm upstart script can just send an event to the xserver upstart script to ensure it's corresponding xserver is running Sep 19 12:31:52 small issue is that.. well there's a reason i've been up all night (7:30am here) so probably won't have any good downtime in the immediate-to-near future, fwiw :) Sep 19 12:32:03 oh, okay. sure thing. Sep 19 12:32:03 ok Sep 19 12:32:13 so the script's job is "ensure there's a display running here" Sep 19 12:32:18 yep Sep 19 12:32:38 and determine that as quickly as possible :) Sep 19 12:32:53 i'll give time initctl a shot when i'm doing it. Sep 19 12:33:02 rwhitby: do you have the exact command you're using handy? Sep 19 12:33:20 "time initctl org.webosinternals.xserver"? or something? Sep 19 12:33:36 i haven't looked too closely at the upstart stuff, and since you were just doing it thought i'd ask :) Sep 19 12:34:00 time initctl start org.webosinternals.xserver Sep 19 12:34:10 (for the current scheme) Sep 19 12:34:18 in the new scheme I'm thinking of, it would change to Sep 19 12:34:38 time initctl emit org.webosinternals.xserver DISPLAY=:X.0 Sep 19 12:34:52 great, ty. Sep 19 12:34:58 maybe even emit org.webosinternals.xserver-start or something Sep 19 12:40:21 yep, nested emits works Sep 19 15:59:30 rwhitby, dtzWill, good job on the new alpha x-stuff :) Sep 19 16:57:42 rwhitby: does the xkb stuff take a significant amount of time, do you know? Sep 19 16:58:15 if so i can work to precompile and integrate it.. pretty sure there's no particularly good resaon to do it on-the-fly every time other than i had (have) no idea how to do it 'right' Sep 19 16:58:36 also zomg yay/good work on all the changes last night :D Sep 19 17:23:01 dtzWill, this will make xapps a success. Sep 19 17:26:08 what did I miss last night? Sep 19 17:28:48 multiple x-server starting and assigning to apps. Sep 19 17:29:50 oh wow Sep 19 17:30:01 excellent Sep 19 18:29:54 I suspect there is a lot of libs that will be common and hopefully only a small bit that isn't able to be re-used Sep 19 19:12:38 what is the current mem usage per X server upon startup? Sep 19 19:17:30 quick, someone launch as many x servers as they can! :) Sep 19 19:17:35 (idk, sorry) Sep 19 19:18:11 * scoutcamper goes to try it Sep 19 19:18:49 * scoutcamper has 10 so far, still going Sep 19 19:19:02 all xterm's of course Sep 19 19:19:57 * scoutcamper has 25 in a stack, TP is lagging now Sep 19 19:20:07 * scoutcamper waits for the TMC to appear Sep 19 19:20:39 dtzWill, got TMC at 30 xterm's Sep 19 19:21:35 :) Sep 19 19:21:54 now, lets see an ipad do that :) Sep 19 21:13:14 we need a tabbed terminal, like mrxvt **** ENDING LOGGING AT Tue Sep 20 02:59:56 2011