**** BEGIN LOGGING AT Wed Jan 25 02:59:56 2012 Jan 25 04:24:00 Brybry, i could send Ben an email about it, im pretty sure he is in charge of all things apis Jan 25 04:45:55 Brybry, dtzWill still awake/around? Jan 25 08:08:25 Brybry: you could try accessing the input device directly Jan 25 08:08:46 i doubt they loose the key events in the kernel Jan 25 08:51:42 stbuehler, luna vacuums up the events and doesn't let them out Jan 25 10:01:06 mmmmmm many good changes :D Jan 25 14:33:40 dtzWill, you are up early, or late Jan 25 14:33:41 cant tell Jan 25 14:33:49 dtzWill, what TZ are you in again? Jan 25 14:34:29 PuffTheMagic: cdt :) Jan 25 14:34:43 so mad early i guess is the right description Jan 25 14:34:57 but my sleep patterns got really really messed up with the NY site visit, 3 nights with 1 hour of sleep and hadn't really had a good chance ot fix since Jan 25 14:35:16 ya i always get messed up when i go to cali Jan 25 14:35:24 once i get back i say up to all hours Jan 25 14:35:35 so yesterday after i did allllll the things (hyperboleandahalf-style) went home around...4pm? chowed down on a burrito and hella nachos and just psased out until this morning, woke up for an hour, and went back to sleep Jan 25 14:35:50 haha, and now all is good in the world again. yay for having a life that lets me do a full reset like that Jan 25 14:36:01 :D Jan 25 14:36:13 today is my first day of school Jan 25 14:36:23 but im still on semester break schedule Jan 25 14:36:28 so im still staying up to 4am Jan 25 14:37:11 dtzWill, im converting wTerm to a headless app now to better handle multi cards and dock mode Jan 25 14:37:31 oh, sweet Jan 25 14:38:06 i had to work very very hard to not start work on gles2-ifying wterm, damn studying :). looks like the rendering is in plenty capable hands but i still am :( that i won't be part of it this time 'round :( Jan 25 14:38:15 PuffTheMagic: what does it mean to be headless, in this context/exactly? Jan 25 14:38:54 it means means that a windowless scene sits around and manages things Jan 25 14:39:26 checks window names (like for the dock) and prevents multiple instances Jan 25 14:39:31 things like that Jan 25 14:39:48 allows for better handling of launch args and shit Jan 25 14:40:12 dtzWill, if you gles2 wterm I need to make a new stub library so that I can build Jan 25 14:40:21 so give me a heads up before you start that Jan 25 14:40:32 preferably after I finish the headless stuff Jan 25 14:40:39 luckily gles uses much much less symbols, generally Jan 25 14:41:05 since beyond basic setup and whatnot (much of which overlaps with glesv1) you hand the gpu some programs and say "run this" :) Jan 25 14:44:04 but yes that can be done for sure, and shouldn't need to be related to anything you're doing immediateyl Jan 25 14:44:28 very glad to see stbuehler's recent changes cleaning much of that up... lots of the 'minor' refactorings we've talked about for a while, all in one commit ^.^ Jan 25 14:45:50 hehe Jan 25 14:46:16 also, good morning to you both :) Jan 25 14:49:44 i think wTerm is a fascinating project Jan 25 14:50:43 javascript, sdl, opengl/shader, utf-8, parsing, terminal handling - so many interesting topics Jan 25 14:51:00 stbuehler, well we are glad to have you Jan 25 14:51:22 agreed. lots of fun technical challenges, lots of things to learn about (how the devil DOES a terminal handle $FOO?! hehe), and it's damn useful to boot Jan 25 14:51:55 ya first term i can actually use on a webos device Jan 25 14:52:15 and a active developer team; that is what keeps my motivation up Jan 25 14:52:22 * dtzWill nods Jan 25 14:52:43 ya having people to work with makes it more fun Jan 25 14:52:48 i have many other interesting projects, but actually working on them is hard as i am often working alone Jan 25 14:53:00 PuffTheMagic, I have used xTerm in the past on the TP... but wTerm is definitely a much more pleasant experience Jan 25 14:53:13 hehe Jan 25 14:53:16 especially the keyboard rocks Jan 25 14:53:31 if only keys didnt get stuck Jan 25 14:53:32 yeah, I wish I could make the built in keyboard be the one from wTerm Jan 25 14:53:38 yeah, minus that Jan 25 14:53:56 once you get the keyboard all dialed in, we should look into hacking it into webOS core Jan 25 14:54:01 perhaps once we get source code... Jan 25 14:54:10 cryptk, when we get the code we can :D Jan 25 14:54:29 cryptk, i was actually aftraid of spending any time on the vkb cause i was sure once i got it the way i liked the source would drop Jan 25 14:54:49 that may still happen, never know Jan 25 14:54:58 code hasn't dropped, wTerm keyboard still has a few quirks Jan 25 14:55:11 could drop the day after you get it fixed, lol Jan 25 14:55:37 just asked my pal murphey, he says that code is likely to drop within 24 hours of your keyboard Jan 25 14:55:58 :D Jan 25 14:56:26 we should be getting enyo any day according to what bheuy said a few weeks ago Jan 25 15:03:39 :( i need to get dressed and head to school :( Jan 25 17:30:14 what up Jan 25 17:37:30 fucking shaders don't do what i want them to... Jan 25 17:38:20 (trying to fill the screen with red right now... glClear() seems to work, could fill it with green that way) Jan 25 19:41:52 bah, I thought all of our lines were initially filled to screen width with blanks now but I guess they're not Jan 25 19:42:59 yep Jan 25 19:43:38 actually, they are on reset Jan 25 19:43:44 maybe it's just insert that they're not Jan 25 19:43:44 hehe, k Jan 25 19:43:54 i think insert make may them even too long Jan 25 19:44:15 aaaargs Jan 25 19:46:53 damn shaders Jan 25 21:05:48 just fyi since I've heard about bip here, there's a buffer overflow in the latest 0.8.8... patch available Jan 25 21:16:45 Brybry: haha tolld you you couldn't count on them being filled :) Jan 25 21:17:12 but they shouldn't be undrefilled, i'd expect insert to create empty ones Jan 25 21:17:21 btw something broke vim for me, which is :( Jan 25 21:17:30 "something" is probably my own changes, ll Jan 25 21:17:42 dtzWill, vim never has worked for me Jan 25 21:17:47 even via novaterm is broken Jan 25 21:17:54 the one installed from optware that is Jan 25 21:18:49 oh, i did my latest set of wterm dev on the tp using wterm+vim locally :) Jan 25 21:19:00 but now scrolling just does bizarre things D: Jan 25 21:19:20 arrow keys dont work for me in vim Jan 25 21:19:56 i feel like installing a funtoo chroot on this TP Jan 25 21:20:03 :D Jan 25 21:20:08 now that i got drobbins to build stage3 for arm :D Jan 25 21:20:49 wonderful. i wsa gonna say i know he was poking at funtoo-izing the tp, either with a chroot or with an actuall install Jan 25 21:21:39 he took my base setup and get it working on metro, their automated stage builder Jan 25 21:21:51 i think when source comes out we need to utilize funtoo/metro for our base Jan 25 21:21:57 and forget what ever palm gives us Jan 25 21:22:01 i really hate OE Jan 25 21:22:26 http://ftp.osuosl.org/pub/funtoo/funtoo-experimental/arm/armv7a/ Jan 25 21:22:41 suppose i need to format and make a ext3 now Jan 25 21:23:08 tailor :D Jan 25 21:23:18 PuffTheMagic: you mean so you can use an even more obscure build system than OE ? Jan 25 21:23:34 ps: I've never actually used that, lol, and prob easier to just do it yourself Jan 25 21:23:39 "easier" Jan 25 21:24:01 portage/emerge is not obscure, and there are nearly everything under the sun linux wise has ebuilds for it Jan 25 21:24:15 making one for luna and hidd and friends will be trivial Jan 25 21:24:20 http://www.yoctoproject.org/ is the obvious choice as far as I can see Jan 25 21:24:57 descendent of OE, supported by the Linux Foundation, lots of SoC vendor support included for easy porting. Jan 25 21:25:00 if it doesnt use uclibc, its not embedded Jan 25 21:25:17 what do we need SoC wise? Jan 25 21:25:29 besides a kernel Jan 25 21:25:34 PuffTheMagic: seriously, you equate uclibc with embedded in 2012 ? Jan 25 21:26:05 i hardly consider a modern tablet an embedded device Jan 25 21:26:06 embedded is multi-core A9 in 2012. pull yourself out of the 20th century. Jan 25 21:26:08 at least not the TP Jan 25 21:27:29 rwhitby, i think its pretty funny you consider gentoo/funtoo obscure and the you link to yocto Jan 25 21:28:01 I think it's pretty funny you link to a desktop build system for embedded work. Jan 25 21:28:28 gentoo isnt "desktop" Jan 25 21:28:32 its what ever you want it to be Jan 25 21:28:58 as if qualcomma and nvidia are going to contribute stuff to funtoo ... Jan 25 21:29:08 s/qualcomma/qualcomm/ Jan 25 21:29:35 again, what would anyone need from them besides kernel support which is distribution agnostic Jan 25 21:30:45 hmm, you think kernel is all you need to use a new SoC for a mobile OS ? Jan 25 21:31:34 yup, as long as the dont make proprietary stacks and new apis (like palm and their bt shit) Jan 25 21:32:29 rwhitby, but you arent answering the question what is an example of something qualcomm would contribute a project like yoctoproject that isnt kernel related? Jan 25 21:33:10 let's see: graphics, video, audio, dsp, etc etc Jan 25 21:33:20 thats all kernel Jan 25 21:33:30 um, no. Jan 25 21:34:30 yay. shader wterm (only ascii, no colors) works. Jan 25 21:34:45 and no cursor :) Jan 25 21:36:10 running cmatrix: WebAppMgr burns at 93%, cmatrix 35%, wterm 18% Jan 25 21:36:25 what's with all bickering, jeepers Jan 25 21:36:30 stbuehler: excellent! Jan 25 21:36:42 stbuehler, thats prob at 384 mhz too Jan 25 21:36:43 well, why is WebAppMgr burning? Jan 25 21:36:45 does the code live anywhere? I'm curious what route(s) you took Jan 25 21:36:47 hm Jan 25 21:37:04 let me push it. PuffTheMagic do NOT MERGE it...! Jan 25 21:37:05 probably b/c we're a plugin and are still pushing through the html5 layer or something? Jan 25 21:38:02 lol Jan 25 21:38:09 i dont merge unless i get requests Jan 25 21:38:10 https://github.com/stbuehler/wTerm/tree/wip_do_not_merge Jan 25 21:38:23 :) Jan 25 21:38:53 i don't have build.git yet, so i didn't pull the latest upstream changes with cmatrix/ncurses Jan 25 21:38:54 stbuehler: lol! Jan 25 21:39:23 now i can start cleaning up the sdlfont stuff Jan 25 21:39:51 i'm not sure how to handle colors yet Jan 25 21:40:23 i could create an extra texture with the color mappings and use another value from the cell texture as index Jan 25 21:40:34 * dtzWill nods Jan 25 21:40:53 pushing the colors (esp when there's only what, 12, of them?!) for every vertex is..unfortunate Jan 25 21:41:58 dtzWill, we could support xterm 256 colors :D Jan 25 21:42:04 if the 12 is an issue ;) Jan 25 21:42:12 and there is 16 Jan 25 21:42:15 8 dim 8 bold Jan 25 21:42:31 and possibly 2 more Jan 25 21:42:33 err Jan 25 21:42:35 4 more Jan 25 21:42:38 for fg/bg Jan 25 21:42:42 dim and bright Jan 25 21:42:47 so max 20 right now Jan 25 21:43:07 PuffTheMagic: 256 color support would be sweet Jan 25 21:45:18 wel, should be possible Jan 25 21:45:23 stbuehler: do you not need to send 4 vertices for each cell still? Jan 25 21:45:26 i use a RGBA texture for cell data Jan 25 21:45:36 each cell has four bytes Jan 25 21:45:50 dtzWill: no, it is one large rectangle Jan 25 21:45:55 (ideally we'd only send the characters, and colors, and shader would 'do the rest'... but i thought you'd still need tooo) Jan 25 21:45:58 ahhh of course Jan 25 21:46:01 exactly :) Jan 25 21:46:33 so fragment shader then combines text data and font glyph table and away we go? Jan 25 21:46:48 yep Jan 25 21:47:24 seems like adding color information would be straightforward, no? throw it in a streaming buffer and consult it while rendering? Jan 25 21:47:36 * dtzWill half-speculation Jan 25 21:48:36 need another texture for the lookup Jan 25 21:49:02 well or tack it on to the character texture? Jan 25 21:49:11 (interleaved with the characters themselves, perhaps?) Jan 25 21:50:11 shrug you built this already, if you have a plan go for it :). just throwing half-baked ideas out there :) Jan 25 21:50:16 :) Jan 25 21:50:23 (b/c i can't justify doing/trying this myself) Jan 25 21:50:28 i think the glyph texture should be luminance only Jan 25 21:50:47 there is no harm in another texture i guess Jan 25 21:51:04 oh can we just bind multiple textures? hahaha Jan 25 21:51:32 i'm so used to using GL as a fast blit that i forgot that was something you could do :3 Jan 25 21:51:50 i just meant interleave the color information with the cell information Jan 25 21:52:24 so you upload structs of {character, color} pairs instead of just characters Jan 25 21:52:28 shrug doens't really matter Jan 25 21:53:15 but mmm glad you're tackling this and looks good :3 Jan 25 22:09:19 OH SHIT! Jan 25 22:09:32 endless term spawning Jan 25 22:09:41 hm? Jan 25 22:09:51 i created a loop by accident :D Jan 25 22:09:54 hehe Jan 25 22:19:42 yay, colors back :) Jan 25 22:23:02 btw, this limits us now to 64 unicode slots :) (fonts * slots <= 256) Jan 25 22:24:43 stbuehler, so how about a way to easily load a custom font and range into a slot Jan 25 22:24:55 and map to a graphics slot Jan 25 22:25:02 not this week :) Jan 25 22:25:05 :D Jan 25 22:25:20 cmatrix is just not the same with out the font :D Jan 25 22:25:36 .. Jan 25 22:25:50 cmatrix is running fine, but mc somehow destroys the graphic state Jan 25 22:25:53 no idea how Jan 25 22:26:04 needs to reinit opengl to get working again Jan 25 22:26:55 cmatrix doest use custom charsets atm Jan 25 22:27:04 mc is the only thing that does Jan 25 22:27:12 and your unicode stuff Jan 25 22:29:18 good idea. that broke it :) Jan 25 22:30:48 i still don't know why though :) Jan 25 22:33:55 it feels like a stupid memory corruption Jan 25 22:56:23 the on-demand glyph rendering doesn't work - if i render them as soon as i have the fonts it works Jan 25 23:40:15 [20120125-18:39:48.114707] error: Uncaught TypeError: Object function (inIterator, inContext) { ^Ifor (var i = 0, len = this.length; i < len; i += 1) { ^I^Iif (inIterator.call(inContext, this[i])) { ^I^I^Ireturn this[i]; ^I^I} ^I} ^Ireturn undefined; } has no method 'replace', src/enyo/plugin.js:159 Jan 25 23:40:17 why **** BEGIN LOGGING AT Wed Jan 25 23:56:08 2012 Jan 26 00:14:12 stbuehler, hows it going over there Jan 26 00:33:37 well i've pushed my headless conversion Jan 26 00:33:44 wterm now runs in dock mode properly Jan 26 00:34:27 i also use ^S / ^Q to suspend/resume cmatrix when docmode is not active Jan 26 00:35:04 im not sure if there is a race condition there Jan 26 00:35:33 i would appreciate it people gave it some stress testing Jan 26 00:36:18 I would if I had a TS for my TP, lol Jan 26 00:36:58 exhibition mode isnt what needs testing Jan 26 00:37:08 the main app does Jan 26 00:37:15 thats when i see the potential race Jan 26 00:37:17 what did you mean by dock mode? Jan 26 00:37:20 well i get some weird error Jan 26 00:37:28 dock=exhibition mode Jan 26 00:37:58 to do proper exhibition mode i needed to convert to a proper headless app Jan 26 00:38:14 there was a lot of refactoring to do that Jan 26 00:38:19 ahh Jan 26 00:38:22 and after I did I started getting some weird bugs Jan 26 00:38:27 but it think i fixed them Jan 26 00:38:39 but just want people to see if they hit any issues Jan 26 00:39:21 gotcha Jan 26 01:35:29 Brybry, what is that gist link of yours Jan 26 01:35:40 nm, got it Jan 26 01:43:43 I think he rotation lock/orientation bug is back in PTM/master Jan 26 01:43:57 i had a feeling about that Jan 26 01:44:02 i noticed you dont subscribe to it Jan 26 01:44:12 i will debug that in a bit Jan 26 01:44:20 i just renabled the wiki Jan 26 01:44:26 and u guys have edit access Jan 26 01:44:34 im moving the input research to it Jan 26 01:45:32 kk Jan 26 01:51:01 https://github.com/PuffTheMagic/wTerm/wiki Jan 26 01:58:37 hmm, I wonder if we can make it pause/sleep when switching to a different exhibition app Jan 26 01:59:21 it already pauses when not active Jan 26 01:59:56 hey Jan 26 02:00:08 ctrl+shit+[ sends escape :D Jan 26 02:00:11 on bt keyboard Jan 26 02:02:02 bt keyboard must really send some junk ouy Jan 26 02:02:13 i dont get why vttest ctrl test fails so bad Jan 26 02:07:37 Brybry, let me know if my last commit fixed rotation lock for you Jan 26 02:11:17 looks good :) Jan 26 02:27:54 why does this not work Jan 26 02:27:55 const char *end = strrchr(argv[0], '/'); Jan 26 02:27:55 end = '\0'; Jan 26 02:28:06 im trying to get the path of the app Jan 26 02:28:40 oh const prob Jan 26 02:28:43 :/ Jan 26 02:33:51 if I set that char to free will there be a memleak when i free it?> Jan 26 02:42:14 I don't think you need to free stuff returned by strrchr Jan 26 02:42:24 before that i did strdup Jan 26 02:42:28 cause the orig was const Jan 26 02:42:31 and i couldnt modify it Jan 26 02:44:06 you do need to free the strdup Jan 26 02:44:10 I think you'd still just free the pointer returned by strdup Jan 26 02:44:22 but not the strrchr one Jan 26 02:45:27 but how does free work? Jan 26 02:45:35 if i set a value mid string to 0 Jan 26 02:45:43 wont it free just up to that point? Jan 26 02:46:25 "man 3 dirname" seems related Jan 26 02:46:49 PuffTheMagic: nopers, free doesn't care what you've written into the array of characters that was allocated previously Jan 26 02:47:34 (but honestly strrchr seems good enough to me) Jan 26 02:47:39 j/w, why do you want the app path? Jan 26 02:47:43 how does it know how big it is? Jan 26 02:47:54 PuffTheMagic: the allocator internals do, yes. Jan 26 02:49:46 oh i thought you asked /if/ it does, i'm sorry Jan 26 02:51:39 i don't actually know how malloc in particular does it Jan 26 02:51:54 but needless to say inside the allocator implementation it tracks all kind of state about the sizes of the various allocations Jan 26 02:53:32 note that malloc has absolutely no idea what you're using is a "string", it just knows allocations and their sizes--that's why malloc takes a size and returns a void* (untyped pointer) Jan 26 02:59:49 ok what am I missing about execvp and PATH **** ENDING LOGGING AT Thu Jan 26 02:59:57 2012