**** BEGIN LOGGING AT Fri Jan 20 02:59:57 2012 Jan 20 03:18:06 [A Jan 20 03:19:05 yos Jan 20 03:19:18 how did your shit go today? Jan 20 03:23:38 eek re [A lol Jan 20 03:24:14 went wonderfully, ty for asking. honestly a bit overwhelmed with how well it went in terms of the feedback and post-presentation discussions and the like Jan 20 03:24:20 although i'm damn exhausted haha Jan 20 03:24:32 passed out 2-6, went to dinner with the group to be social and am exhausted still Jan 20 03:24:42 would crash 'cept i wanna be sleepy for the flight back tomorrow :) Jan 20 03:24:50 esp since the plane won't have internet Jan 20 03:25:03 did you guys sort out the emacs issue? Jan 20 03:25:20 (and j/w what's the vi bug via ssh? :)) Jan 20 03:26:21 ya emacs issue is fixed Jan 20 03:26:27 currently working on the erase bug Jan 20 03:26:46 i dont fully understand the vi over ssh bug cause i dont get it Jan 20 03:26:49 PuffTheMagic: multi-card wterm is critical for the end goal of being able to dev wterm on wterm  Jan 20 03:26:50 but there is def issue with ssh Jan 20 03:27:02 dtzWill, multi-card wterm is done Jan 20 03:27:05 and already released Jan 20 03:27:36 sorry i'd have alredy written "so thank you for doing that" if i wasn't lagging like a madman Jan 20 03:27:47 apparently madmen are laggy as shit, in case you didn't know. Jan 20 03:27:54 :D Jan 20 03:27:55 * dtzWill finds phone charger so he can tether Jan 20 03:28:11 the multi card wterm was like 2 lines of code Jan 20 03:28:22 using freetether :D Jan 20 03:28:28 :D Jan 20 03:28:42 looking at the erase code I see is has more issues than I though Jan 20 03:28:56 he actually removes cells in the buffer Jan 20 03:29:03 instead of replacing them with spaces Jan 20 03:32:17 all the operations need to ensure that the buffer lines have a min length of at least the width of the term Jan 20 03:36:16 so here's a question Jan 20 03:36:22 why don't we just define the screen as a grid Jan 20 03:36:30 and to hell with this nonsense Jan 20 03:36:34 serious question Jan 20 03:36:34 lol Jan 20 03:36:40 resizing? Jan 20 03:36:50 well right so then you size the grid Jan 20 03:37:31 but okay what i'm getting at is it seems simpler if the code never had a nothing character for a cell Jan 20 03:37:34 thats basically what we have now Jan 20 03:37:42 since what we're suggesting is erasing to blanks Jan 20 03:37:42 okay Jan 20 03:38:10 its not that a cell has a nothing character Jan 20 03:38:18 just that invariants like that makes things a lot easier to reason about Jan 20 03:38:22 its that the "term" is a vectors of vectors Jan 20 03:38:31 so the length of a line varries Jan 20 03:38:37 well sure, we put nullbytes there to indicate it's unused when erasing or when it's not initialized Jan 20 03:38:40 well it *can* vary Jan 20 03:38:55 there's no reaso we don't prepopulate it with ' 's and just never ever put a nullbyte there Jan 20 03:39:17 i imagine that optimization was much more important b/c of the slow rendering Jan 20 03:39:54 well anyway i'll save ideas like that for another time perhaps, but the important thing i was trying to get at is that there's not externally visible difference between a cell that has a blank and a null-byte Jan 20 03:40:03 again, its not like thre are null bytes Jan 20 03:40:07 there are no null bytes Jan 20 03:40:15 the length of the line is changing Jan 20 03:40:59 when a char is deleted at location x,y a space should be appended at the end of the line Jan 20 03:41:19 and when a char is erased it should be erased with a space Jan 20 03:41:26 hmm Jan 20 03:41:29 right but is the notino of 'line length' an actual thing terminals do Jan 20 03:41:40 but then again, there is a separate vector for the graphics states Jan 20 03:41:48 or just we happen to not use the full line until some operation goes over there, after which we erase to blanks Jan 20 03:41:57 not sure why there are 2 variables, 1 for text and 1 for gs Jan 20 03:43:46 (idk) Jan 20 03:48:46 wonder what's wrong with the ipkg vim. Jan 20 03:48:59 maybe building my own would work better? (misc changes, etc....) Jan 20 03:49:25 ya i was considering that today to Jan 20 03:50:03 when i ssh into my desktop and use the vim there is perfect Jan 20 03:50:47 dtzWill, we should probably package it up for preware Jan 20 03:50:53 get it building in build.git Jan 20 03:58:55 yay it built on device Jan 20 03:58:57 * dtzWill verifies it actually runs Jan 20 03:59:45 hmm does wterm have any tricks in postinst? or can i just clone the app dir and modify the appinfo.json? Jan 20 03:59:58 looking to have a test version since i can't very well overwrite my currently running one Jan 20 04:00:04 if i'm using wterm to hack on wterm Jan 20 04:01:19 just a postint to add vttest to path Jan 20 04:01:43 hahaha it runs fine Jan 20 04:01:52 had to run it with pwd set to the appdir Jan 20 04:01:56 but still Jan 20 04:02:08 err and idk if i needed to but dropped the napp_plugin_id, and LD_ values from my env Jan 20 04:04:08 interesting result: control-c is eaten when run standalone Jan 20 04:04:16 as are control-v, and control-a Jan 20 04:06:02 what was those last 2 lines about Jan 20 04:10:50 oh to run it in non-plugin mode from a wterm instance Jan 20 04:11:19 i dropped some things from my env, idk if i needed to (tried it before thinking to change my working directory to the app dir, since that's where we expect to load resources from using relative paths) Jan 20 04:11:32 and then when launched that way control-c and control-v don't work Jan 20 05:09:33 anyone know how to use irssi with bip? Jan 20 05:10:29 bip? Jan 20 05:11:12 * rwhitby uses bip, but not irssi Jan 20 05:11:46 or znc Jan 20 05:11:55 i dont get how to get irssi working with a bouncer Jan 20 05:28:40 oh Jan 20 05:28:54 I use irssi as a "bouncer" but that's it Jan 20 05:49:46 got bip working in irssi Jan 20 05:50:18 how do i scroll up/down Jan 20 05:53:39 * PuffTheMagic wishes he could scroll Jan 20 06:43:39 PuffTheMagic: /sb goto -20 Jan 20 06:43:46 PuffTheMagic: /sb goto +20 Jan 20 06:44:12 /help scrollback Jan 20 06:54:09 PuffTheMagic: do you have an esc key? Jan 20 06:57:00 PuffTheMagic: can you explain the logic used for the key handling (esp control, etc)? do we have a source for that? Jan 20 07:00:45 logic Jan 20 07:00:52 ? Jan 20 07:02:11 PuffTheMagic: the code in the braces for the condition "if ( mod & KMOD_CONTROL)" Jan 20 07:02:33 so many magic numbers :) Jan 20 07:03:10 trying to make it work for other things, just wondering what inspired that encoding. Jan 20 07:03:39 well it would be simpler if sdl had syms for uppercase Jan 20 07:03:53 then they all woulc be -64 Jan 20 07:05:29 what doesnt work? Jan 20 07:06:13 PuffTheMagic: if you have an esc handy, but no alt, esc,p and esc,n can go up/down buffer (like pgup/pgdn), or alt-p/alt-n (in irssi) Jan 20 07:06:52 i have alt and esc Jan 20 07:08:52 ahhh Jan 20 07:09:19 the prob is that scrolling is broke for xterm Jan 20 07:09:28 works for rxvt Jan 20 07:16:02 dtzWill: so what is the issue with ctrl Jan 20 07:19:18 dtzWill, http://www.asciitable.com/ Jan 20 07:19:35 ^^ that should help u understand what is happening with ctrl Jan 20 07:19:48 if ctrl+ anything in the 3rd column is pressed Jan 20 07:20:03 its -64 Jan 20 07:20:22 but SDL doesnt have sym codes for upper case just lower case Jan 20 07:20:41 so when A is pressed u get sym 97 instead of 65 Jan 20 07:30:58 PuffTheMagic: bit why do you shift by 64 when control is pressed? Jan 20 07:31:31 i mean i understand *what* it does, my question is 'why'. is there a particular standard control-modifier encoding that adheres to? Jan 20 07:32:40 all i found was a leonerd.org.uk/hacks/fixterms page Jan 20 07:33:24 anyway i don't know, i was just curious. Jan 20 07:33:33 what doesn't work is control+"various things" Jan 20 07:33:41 where "various things" i've tried is '\' and '/' Jan 20 07:34:03 and so i wanted to understand how the existing stuff worked :) Jan 20 07:34:40 vttest Jan 20 07:34:48 thats all i can say Jan 20 07:35:37 okay, so you reverse engineered the encoding Jan 20 07:35:49 basically Jan 20 07:35:51 based on when the lights went off on vttest? or from it's matching source code? Jan 20 07:35:58 kk Jan 20 07:36:08 well that's damn useful for verifying any changes i make anyway Jan 20 07:36:49 i mean i think it makes sense Jan 20 07:37:08 but ive seen no offical docs about ctrl bindings Jan 20 07:47:42 dtzWill: what dont work that u r trying to fix? Jan 20 08:04:49 well control-\\ Jan 20 08:04:52 err control-\ Jan 20 08:05:35 but also figure out how to do it 'right' regardless instead adding a one-off again Jan 20 08:08:19 fwiw Jan 20 08:08:46 http://www.leonerd.org.uk/hacks/fixterms seems useful Jan 20 08:09:04 while not authoritative, perhaps, certainly useful Jan 20 08:10:11 also shows how we might insert arbitrary utf8 characters, with modifier support Jan 20 08:19:19 merged Jan 20 08:21:08 zomg thank god for optware Jan 20 08:21:19 ? Jan 20 08:21:33 ipkg-opt install coreutils zsh less grep ..... etc Jan 20 08:21:45 so not stuck with crappy busybox versions Jan 20 08:22:13 gottapackage up vim Jan 20 08:25:07 vim seems to work fine for me atm Jan 20 08:25:16 i think not using busybox sh helps Jan 20 08:25:22 but that's a tad unfounded Jan 20 08:25:41 i also dropped those LD_* env vars too which might've helped Jan 20 08:26:01 not saying it's advice per se, but now that you mention it it seems to be fine now Jan 20 08:26:19 well other than we don't handle all its codes properly quite yet Jan 20 08:29:21 latest wterm shoudlnt have those in their env Jan 20 08:30:25 oh, okay Jan 20 08:30:32 and vim is still broke for me Jan 20 08:30:42 even using a non-lame shell? :( Jan 20 08:30:55 bash Jan 20 08:32:55 real bash, not busybox bash? Jan 20 08:33:06 bummer Jan 20 08:33:09 what escape codes are u seeing in vim? Jan 20 08:33:14 ya real bash Jan 20 08:34:52 PuffTheMagic: weirdness when scrolling Jan 20 08:35:03 text not being updated right, and what looks like stale graphicstates Jan 20 08:35:09 (not sure, clearly) Jan 20 08:35:19 looking into erase thing first Jan 20 08:35:47 if you're bored could you try vim with zsh? it honestl is working fine for me, complete with my own .vimrc and such Jan 20 08:37:45 ya scrollig is not right Jan 20 08:39:29 zsh is horribly broke on wterm Jan 20 08:39:43 and vim still broke Jan 20 08:49:57 bummer Jan 20 08:50:15 vim working here crash-wise, idk what that's about Jan 20 08:50:40 crash wise? Jan 20 08:50:53 what was broke about it for u/ Jan 20 08:53:37 oh previously it'd instacash Jan 20 08:53:39 crash Jan 20 08:53:49 ah Jan 20 08:54:00 never craahed for me Jan 20 08:54:08 but cursor keys are fucked Jan 20 08:54:47 insta-crash? hmm, wonder if it crashes if you re-set LD_PRELOAD ... Jan 20 08:55:20 it never crashed for me Jan 20 08:55:36 even before we removed -p Jan 20 08:56:28 guess I should install vim Jan 20 09:07:25 * dwc- needs to start assigning static leases to his webos devices Jan 20 09:07:56 my touchpad is either not on my wifi, or sshd isn't answering Jan 20 09:08:19 * dwc- puts off vim install until it doesn't involve an extra trip out of bed Jan 20 09:09:06 dwc-: :D Jan 20 09:32:56 PuffTheMagic: re: the erase bug, any more test cases besides htop and that escape sequene you gave me? Jan 20 09:33:13 (i believe it's fixed) Jan 20 09:43:00 PuffTheMagic: do you know what escapes are used for scrolling? Jan 20 10:01:19 hmm sees vim is using delete lines Jan 20 11:52:54 dtzWill: this is what I've seen of bt key differences/issues so far https://gist.github.com/1646976 Jan 20 11:53:38 amazingly none of the alt key combinations have issues Jan 20 16:11:47 dtzWill, no pull request? Jan 20 16:12:04 dtzWill, well i gave u 6 pull requests Jan 20 16:13:35 lol Jan 20 16:13:55 now i see it Jan 20 16:14:03 nope Jan 20 16:14:10 that was the Control- one Jan 20 16:38:30 dtzWill, just merged your fix for htop Jan 20 16:38:46 i cant believe what you did is all it took Jan 20 16:39:00 i guess i dont understand the graphics state code in this app Jan 20 16:39:03 however Jan 20 16:39:12 character attributes need to be reset Jan 20 16:39:21 like underline and blink Jan 20 16:39:36 the echo test still fails Jan 20 16:45:34 fails 2 echo tests Jan 20 16:45:57 dtzWill, echo -e -n "\e[2J\e[20;20H\e[1;4;5;44mHello\e[J\e[m\n" Jan 20 16:46:02 it doesnt pass that yet Jan 20 16:51:08 actually i like this test better echo -e -n "\e#8\e[20;20H\e[1;4;5;44mHello\e[J\e[m\n" Jan 20 16:51:14 i will put it in the bug report Jan 20 16:52:45 any progress with utf-8? Jan 20 16:53:19 nope, im putting it on the backburner till people who need it give more details about its behavior Jan 20 16:57:22 damn Jan 20 16:57:24 well Jan 20 16:57:38 i'm not sure how to handle drawing the characters Jan 20 16:58:16 to be honest, i though rendering was already fast before dtzWill added the glyphcache and implemented all the optimizations Jan 20 16:58:17 the texture mapping probably won't work with 2^16 chars... Jan 20 16:59:12 some basic utf-8 decoding to codepoints is not that difficult Jan 20 17:42:21 just "fixed" the erasing behavior Jan 20 17:42:24 pushed Jan 20 17:42:30 needs a sanity check! Jan 20 17:44:46 https://github.com/PuffTheMagic/wTerm/commit/0db4a834103c92b7bd3acbb870d5bfe9611fe5b3 Jan 20 17:55:26 terminal\terminalstate.cpp:293: error: invalid conversion from 'int' to 'TSCharset_t' Jan 20 17:56:34 hmm Jan 20 17:56:44 this is what I get for being on windows and not using the same makefile Jan 20 18:05:18 easy fix Jan 20 18:05:30 but besides that my commit was sorta invasive Jan 20 18:05:35 idk how much else i broke Jan 20 18:05:40 i also removed a bunch of code Jan 20 18:05:50 so that code was either useless or i broke something else ;) Jan 20 18:07:09 Brybry, just pushed a fix for that issue Jan 20 18:07:46 eww Jan 20 18:08:05 but i do notice an existing bug now happens more often Jan 20 18:08:20 and anyone ever start wterm and all they get is a black screen and cursor in top right Jan 20 18:08:28 as if a redraw was skipped or something Jan 20 18:10:53 at first I thought the new removeGraphicsState function was some crazy illogical recursive function :D Jan 20 18:10:57 yeah, sometimes Jan 20 18:11:25 I feel like it's SDLWaiting when it does that Jan 20 18:11:40 but then it doesn't entirely make sense to me as the blink thread should keep it fairly awake Jan 20 18:11:43 Brybry, we might need an initial SDL_Pump or something Jan 20 18:16:35 Brybry, so u said at first... does it make more sense now? Jan 20 18:19:38 why does our event loop use SDL_WaitEvent and SDL_PollEvent? Jan 20 18:19:39 looks like it should work fine to me though maybe slightly inefficient Jan 20 18:19:51 SDL_WaitEvent blocks/is really efficient idling Jan 20 18:19:58 right.... Jan 20 18:20:01 it's better than SDL_Delay even from my understanding Jan 20 18:20:02 but why the polling? Jan 20 18:20:23 to clear out the event stack Jan 20 18:20:37 or are you asking why not loop with SDL_WaitEvent or something? Jan 20 18:20:45 ya basically Jan 20 18:21:07 but i get why Jan 20 18:21:12 never mind Jan 20 18:22:22 if removeGraphicsState is going to get called a lot we might want to write it a tiny bit more efficiently but it's not a big deal either way Jan 20 18:22:35 oh im sure it can be more efficient Jan 20 18:22:46 don't really need the first line and last line checks in there every loop Jan 20 18:23:05 small potatoes though Jan 20 18:25:33 well this last change has actualyl caused the no render bug to happen every time Jan 20 18:27:46 Brybry, if the graphics state vector was changed to be similar to the data buffer vectors Jan 20 18:27:59 when we could do more batch operation for setting the graphics state Jan 20 18:36:54 oops Jan 20 18:37:02 i think i see how i could have simplified something Jan 20 19:00:04 i dont understand c++ iterators Jan 20 19:17:30 fuck it Jan 20 19:17:40 someone who undersands c++ can optimize this Jan 20 21:20:15 PuffTheMagic: sent a pull request for orientation fix stuff again, I missed a case (if someone set rotation lock while the application was running it could lead to the wrong keyboard showing) Jan 20 21:21:21 already merged Jan 20 21:22:16 i got the just type action working Jan 20 21:22:16 sorta Jan 20 21:22:20 there is a race Jan 20 21:22:23 which im trying to make not a race Jan 20 21:24:05 but i cant figure out how to make it not a race because we have to wait for something in a forked child process Jan 20 21:44:41 testy test http://www.ustream.tv/channel/scale-10x Jan 20 21:48:08 did you bring the windows laptop again Jan 20 21:48:43 taht social stream is disturbing, btw... I hope that's some stock thing Jan 20 21:49:38 it is stock Jan 20 21:49:42 and windows _desktop_ Jan 20 21:51:05 now I'm bored Jan 20 21:51:14 and that is a very sensitive mic Jan 20 21:52:03 y'all are boring Jan 20 21:52:12 just setting stuff up Jan 20 21:53:25 halfhalo, u finish that slideshow? Jan 20 21:53:39 ya Jan 20 21:56:16 looks like it hasnt changed Jan 20 21:56:31 aha Jan 20 21:57:16 halfhalo, this show is boring, smile for the camera Jan 20 21:57:30 camera is pointed away from me atm Jan 20 21:57:40 walk to the other side Jan 20 21:57:43 camera angle looks a little high anyways Jan 20 21:57:45 I'm setting something else up atm Jan 20 21:57:55 and it is a tad high, but it works for viewing people Jan 20 21:58:18 thats a busy conference Jan 20 21:58:20 that's true. we want to see the booth babes, not you anyways Jan 20 21:58:32 only setting stuff up Jan 20 21:58:49 expo wasn't open the first day last year, iirc Jan 20 21:58:54 yup Jan 20 21:59:19 HP got a booth again this year? Jan 20 21:59:25 yup Jan 20 21:59:36 I guess all they can raffle off this year is touchpads Jan 20 21:59:42 mebe Jan 20 22:00:37 I see a halfhalo Jan 20 22:00:41 me to Jan 20 22:00:46 nice guns Jan 20 22:01:36 lol Jan 20 22:02:29 who are these ganstas? Jan 20 22:02:30 say hi to bem for me Jan 20 22:02:41 and whoever the boxee guy is Jan 20 22:02:57 lol Jan 20 22:04:51 these ads are awesome Jan 20 22:05:07 do channel popout, then there's no ads Jan 20 22:05:33 either that, or adblock is taking care of me Jan 20 22:06:04 adblock is taking care of you Jan 20 22:07:54 ... Jan 20 22:09:04 off air? Jan 20 22:09:05 wtf Jan 20 22:12:35 bacxk Jan 20 22:13:02 halfhalo, WOSI=iOS+android? Jan 20 22:15:22 I feel like I should just have you put in the key repeat stuff and then someone else can make it better/decide if it's better off in its own thread or what Jan 20 22:15:45 im afraid of the keyrepeat while we have the stuck key issue Jan 20 22:16:14 the web browser Jan 20 22:16:15 the web browser Jan 20 22:16:17 the web browser Jan 20 22:16:34 mmm Jan 20 22:16:39 the key stuck issue with key repeat is not a problem Jan 20 22:16:54 how is it not? Jan 20 22:16:56 at least, I don't think it is Jan 20 22:17:06 it wasn't last time I tried it Jan 20 22:17:08 blackberrys browser is meh, slightly better than webos. Jan 20 22:17:27 dwc-, who is this guy? Jan 20 22:17:37 halfhalo, ^^ Jan 20 22:17:47 friend of mine Jan 20 22:17:54 works for snapfish Jan 20 22:18:36 PuffTheMagic, I think it's not a problem because the stuck key issue still requires a key up to happen on other keys which resets the key repeat Jan 20 22:19:08 requires a key up? Jan 20 22:19:14 he "dragged" me to scale9x last year Jan 20 22:19:25 yeah, I have an 'H' stuck down right now but no repeating H Jan 20 22:19:34 I dunno how to say it Jan 20 22:19:51 can you get a stuck key without hitting multiple keys/firing off a key up for at least one key? Jan 20 22:20:43 idk, send me the request Jan 20 22:20:47 i can always murder you later Jan 20 22:21:52 so where is the updated slideshow? Jan 20 22:29:28 I found something where the stuck key bug sucks =( Jan 20 22:29:30 shift key Jan 20 22:29:43 but I don't think that's key repeat related Jan 20 22:29:52 ya the stuck key bug sucks for sure Jan 20 22:30:51 I was playing around with drag events and setting keys down on dragstart and stuff and it seemed to help some cases but I think it was just treating the syptom and not the problem Jan 20 22:30:59 we'll figure it out one day Jan 20 23:07:29 I love that I can launch multiple wterms and navigate amongst them and dismiss them without ever actually touching my touchpad Jan 20 23:12:43 :D Jan 20 23:49:49 I think we need to tell someone at HP about them breaking ctrl+a/c/v (and probably the other 'eaten' ctrl codes) and maybe ask them if they can still pass along the events after doing their copy/paste/select stuff Jan 20 23:49:57 and then pray and wait a very long time for the next patch Jan 21 02:29:12 Brybry, ya it should be a selectable mode for the keyboard Jan 21 02:29:15 not something forced Jan 21 02:29:56 just like text input boxes have ways of setting behavior of typed text Jan 21 02:30:00 like auto correct, etc **** ENDING LOGGING AT Sat Jan 21 02:59:57 2012