**** BEGIN LOGGING AT Thu Jan 19 02:59:57 2012 Jan 19 10:57:15 dtzWill, I think sym does send alt but I'm not sure that there are any keybindings set up for it Jan 19 15:00:53 Brybry: kk. when i get back i'll give it a look :) Jan 19 15:19:24 dtzWill, multiple cards is on my todo list Jan 19 15:19:28 or tabs or something Jan 19 16:05:54 tabs would be sweet. could we just add a 'tab bar' in the top...10 pixels Jan 19 16:06:21 well nvm i don't mean to try to design this correctly right now Jan 19 16:06:46 i just was thinking that having distinct plugin instances would be overkill, and we could probably do a good deal better Jan 19 16:07:03 via proper separation of our rendering framework and the data itself Jan 19 16:07:11 yay model/view/etc Jan 19 16:08:14 i guess multiple cards might just be an easy and clean solution, shrug Jan 19 16:22:43 Brybry: similar situation re:arrow keys? Jan 19 16:24:52 https://github.com/Brybry/wTerm/blob/experimental/src/plugin/sdl/sdlterminal.cpp#L569 <-- arrow keys Jan 19 16:25:25 different syms depending on plugin or not =( Jan 19 16:48:54 h Jan 19 16:49:10 i was just rewriting the key handling, with next step being adding various bt things Jan 19 16:49:15 * dtzWill looks at changes Jan 19 16:54:16 can definitely just grab those specific syms, we don't need to know that the bt keyboard is attached and all Jan 19 16:54:44 I did that more for completeness/because I could/making the vkb hide on bt keyboard detection Jan 19 16:55:23 I don't think it's worth the fact that you have to link like 5 more libs =/ Jan 19 16:57:16 yeah i already grabbed those same syms from x server Jan 19 16:57:39 althoough wasn't thinking would bother supporting the non-plugin case, is that an important use case? Jan 19 16:58:03 also re:hiding vkb i was thinking of adding support for the 'dismiss' event, and just do it that way Jan 19 16:58:17 only works for hp kb though i suppose Jan 19 16:58:32 (but still a reasonable thing to support regardless, i imagine) Jan 19 17:19:22 Brybry: oh i see re:key codes, i'm sorry i misread Jan 19 17:37:22 oh hey alt works on master Jan 19 17:37:23 hooray Jan 19 17:37:24 irssi :D Jan 19 18:07:38 hmm, bummer that webos still om nom nom's control+{a,x,c,v} :( Jan 19 18:07:44 from the bt keyboard i mean Jan 19 18:08:01 at least it's hooked up usefully to copy/paste in 3.0.5 (which is sweet!) but still :( Jan 19 18:08:09 pass it to the sdl apps anyway, guysssss Jan 19 18:14:29 * Brybry bubbles dtzWill Jan 19 18:18:23 Brybry: bubbles? Jan 19 18:18:25 lol Jan 19 18:18:51 not being able to effectively use screen is rather.. unfortunate :( Jan 19 18:19:06 guess i could switch the control char from control-a to controrl-\ or something.... Jan 19 18:19:14 although control-a in vim is useful too O:) Jan 19 18:19:17 oh wait Jan 19 18:19:17 hmm Jan 19 18:19:43 I would swear ctrl+a worked for me in screen in 3.0.5 Jan 19 18:19:45 but maybe not Jan 19 18:20:00 Brybry: fwiw issued pull req for adding bt arrow keys Jan 19 18:20:10 doesn't seem to here :( Jan 19 18:27:29 dtzWill, what about ctrl+alt+f# to switch vt's Jan 19 18:30:47 damn, you're right Jan 19 18:30:48 they did break that Jan 19 18:36:46 dtzWill, I'm not 100% sure of this but I would bet it's just a plugin issue Jan 19 18:37:19 and we can maybe catch the copy/paste/select bit and then pass the proper stuff to the terminal plugin Jan 19 18:43:10 Brybry: it's not just a plugin issue Jan 19 18:43:15 it happens for all sdl apps Jan 19 18:43:21 have this problem in x/x-server Jan 19 18:43:26 that's why the modifiers there are sticky Jan 19 18:44:27 PuffTheMagic: what about ctrl-alt-f#? and i wouldn't expect that to work, although that's mostly b/c i'm not sure how that normally works Jan 19 18:44:28 =( Jan 19 18:44:55 Brybry: however maybe the enyo portion can hijack the copy/paste things? maybe there are events or something? :/ Jan 19 18:45:28 PuffTheMagic: you mentioned the eraseline bug, should all erase variants retain the original graphics state for the erased cells? Jan 19 18:46:41 dtzWill, i didnt mean real vt switching Jan 19 18:47:00 but that could be the keybinding to switch tabs Jan 19 18:48:38 oh, i see haha :). well that sounds fine except i love using my bt keyboard and it has a bunch of media keys instead of those :( Jan 19 18:49:27 what about if other ttys showed up in the app menu Jan 19 18:49:51 might even be able to allow the term to set the names that show up there Jan 19 18:49:58 or something like other terms Jan 19 18:50:20 %d : %n : %w Jan 19 18:54:44 dtzWill, so with the commit i just merged Jan 19 18:55:06 that breaks like 4 ctrl+ combinations Jan 19 18:55:49 Ctr+S|T|U|V Jan 19 18:56:19 okay, i'll fix. Jan 19 18:56:31 sorry, since it was undocumented magic i wasn't sure how to test and figured you'd have an idea Jan 19 18:56:55 well wait a sec Jan 19 18:57:00 since this is a plugin app Jan 19 18:57:11 we dont need all 8 of those new defines Jan 19 18:57:20 only the 4 that are out of the ascii range right? Jan 19 18:59:35 errrr 'sym' isn't ascii, right? Jan 19 18:59:58 it mostly is Jan 19 19:00:00 well Jan 19 19:00:05 or do we encode ascii chars in the 'sym' field in our key event generator? Jan 19 19:00:25 but i mean at no point do we treat it directly--we use unicode for the non-special keys, maybe? Jan 19 19:00:38 http://www.koders.com/c/fid8D6280AEFC1089A8F98EB34B12BA678E634E5C6A.aspx Jan 19 19:01:50 well if I do a ctrl+S, wterm sends \x13 Jan 19 19:01:56 which is 19 Jan 19 19:02:14 which would then cause wterm to interpret that as a cursor kew Jan 19 19:02:20 from the looks of the new code Jan 19 19:02:31 PuffTheMagic: i know they /happen/ to do so, but do we ever take advantage of that? Jan 19 19:03:00 and as such why do we care that we have these sdl keys that don't follow that convention? Jan 19 19:03:01 what, ctrl bindings? Jan 19 19:03:26 i'm talking about the use of 'sym', and it's comparison regarding those 8 defines Jan 19 19:03:55 i thought I just explained why 4 of them might be an issue Jan 19 19:04:07 while dropping the non-plugin ones would work, why do so when apparently brybry likes running wterm in non-plugin mode Jan 19 19:04:13 HP_BT_PLUGIN_& defines are fine Jan 19 19:04:29 are you saying the control+s bug is related to the other 4 defines? Jan 19 19:04:51 seems that way Jan 19 19:06:12 okay i understand you better now, hmm. Jan 19 19:06:37 when i try ctrl+s it seems to lock up the term Jan 19 19:06:40 idk why thats happening Jan 19 19:06:54 i just though it would cause a cursor event Jan 19 19:07:27 dtzWill, and I think Brybry only uses wterm in non-plugin mode for debugging Jan 19 19:07:33 sonofa Jan 19 19:07:36 1 sec sorry Jan 19 19:08:11 wait Im not even running your merge yet Jan 19 19:08:19 i just realized Jan 19 19:08:26 LOL i scrolled up in irssi Jan 19 19:08:38 and with neither xterm nor wterm can i seem to figure out how to scroll back down Jan 19 19:08:59 well that fixed that >_> Jan 19 19:09:12 PuffTheMagic: i'm sorry, would you mind paste(bin?)ing what you said? Jan 19 19:09:16 i saw you pinged me but couldn't read it Jan 19 19:09:26 control-s is /supposed/ to lock up your terminal btw Jan 19 19:09:32 unlock with control-q Jan 19 19:09:36 try it on any other terminal Jan 19 19:09:44 well lock the shell anyway Jan 19 19:09:50 lol Jan 19 19:09:50 it's flow-stop or something Jan 19 19:10:05 i pinged you? Jan 19 19:10:12 i thought so, maybe not. Jan 19 19:10:25 well okay or don't paste if it's not important :) Jan 19 19:10:33 i just said Brybry probably only uses wterm in non plugin mode to debug with gdb Jan 19 19:10:41 oh, okay Jan 19 19:10:49 well i don't see how it would break anything Jan 19 19:11:35 since afaict the 'sym' field is used to identify SDL key symbols, and any connection to ascii is cute but it's not like our code ever uses 'sym' as an ascii field such su Jan 19 19:11:50 dtzWill: try escape (alt/meta/whatever) n and p Jan 19 19:12:03 or if you lack those, /sb end will scroll to the bottom Jan 19 19:12:08 ah /sb end Jan 19 19:12:17 i...i'm shamefully unfamiliar with /sb Jan 19 19:12:20 tyvm Jan 19 19:12:58 dtzWill, i guess you are right, that that code is only for input generated by keyboard and wont be affected Jan 19 19:13:04 /sb goto is also useful but I can never remember the args Jan 19 19:13:09 PuffTheMagic: sorry i'm not trying to be difficultl, i want to fix any bugs in the code i sent you. does anything not work atm? Jan 19 19:13:24 dwc-: yeah all i remember is that using /sb i could more easily jump to where i was pinged or something Jan 19 19:13:31 but...never used it enough to like it Jan 19 19:13:38 brb bouncing luna b/c it decided to perma-show the vol widget Jan 19 19:13:41 throwing out the 19/non-plugin keybinds is fine tbh Jan 19 19:13:58 that's kind of the whole reason I'd never made a branch/pull request for any of that Jan 19 19:14:17 it seems to be fine Jan 19 19:14:24 im just out of it Jan 19 19:14:46 i went to rochester yesterday and brewed beer at my buddies brewery all day Jan 19 19:14:54 from ~10 am to 10pm Jan 19 19:14:58 and we drank that whole time Jan 19 19:15:02 and then went out after Jan 19 19:15:03 lol Jan 19 19:15:04 :D Jan 19 19:15:10 just drove back this morning Jan 19 19:15:17 and passed out for like 2 hours Jan 19 19:15:24 i've only had 1/2 a coffee so far Jan 19 19:15:36 im usually almost done with a pot by this time of day Jan 19 19:15:59 so since the keycode shit seems to be ok im gonna go back to making scroll buffer work Jan 19 19:24:08 PuffTheMagic: i flew into rochester yesterday Jan 19 19:24:13 up at cornell atm Jan 19 19:24:20 and flying out of rochester tomorrow Jan 19 19:25:48 PuffTheMagic, Brybry: well sounds like maybe non-plugin use case is even less important than we thought, althoough shouldn't hurt anything. i'm fine with whatever y'all wanna do. Jan 19 19:26:11 so u driving to rochester at some point tomorrow? Jan 19 19:26:24 well my coworker drew the short strwa on that one Jan 19 19:26:29 but yep, ochester airport anyway Jan 19 19:26:37 brb a change i made poking at the ersae is messing things up i think Jan 19 19:26:38 haha Jan 19 19:26:42 should stop by Roc Brewing Co Jan 19 19:26:53 have a wterm hackup Jan 19 19:27:26 http://rocbrewingco.com/ Jan 19 19:27:38 i helped make the Belgian Blond yesterday Jan 19 19:42:37 wow Jan 19 19:43:59 wow what? Jan 19 19:44:21 oh that you made a beer they're selling idk Jan 19 19:44:38 well u wont drink the batches i helped make Jan 19 19:44:42 most people who brew stuff, the result ends up in the fridgerator for private consumption Jan 19 19:44:42 and it wasnt my recipe Jan 19 19:44:59 it just went out to hang out for the day and help him Jan 19 19:45:29 but ya he doing well for being in business for only 7months Jan 19 19:45:50 ~14 bars around rochester 1 or more of this beers on tap Jan 19 19:46:05 and he had a little tap room in his brewery too Jan 19 19:46:26 the behavior you described for eraseline (that's broken), the rest of the the 'erase' control characters from the xterm that etla.org link (no paste ability atm).... should they all be 'clear $FOO' instead of 'erase $FOO'? Jan 19 19:46:51 errr well lemme rephrase. there's some operation 'A' that /removes/ characters and their graphicstate. Jan 19 19:47:02 there's a different operation that blanks the characters but keeps their state Jan 19 19:47:26 'B' Jan 19 19:47:43 you said we need to do 'B' for eraseline, is that true of any of the others? Jan 19 19:48:31 well what seems like the behavior actually implemented is a third option Jan 19 19:49:10 characters are erased but a whole new graphicstate is set for the erased cell Jan 19 19:49:45 okay, right, we need to do 'B' for eraseline, but we're not doing that (or 'A'), if i'm understanding right. Jan 19 19:49:58 should all the erase* escape sequences be 'B' then? Jan 19 19:50:57 im not agreeing that A or B should be done Jan 19 19:51:03 i know what the vt102 docs say Jan 19 19:51:10 and then there is what xterm does Jan 19 19:51:11 lol Jan 19 19:51:19 sorry i just know you've spent a non-trivial amount of time with these and know this stuff much better than me Jan 19 19:51:26 PuffTheMagic: oh, i see. okay, hrrm. Jan 19 19:51:31 what should wterm do, then? :) Jan 19 19:51:41 what xterm and most other terms do Jan 19 19:51:46 (clearly what we do now isn't right, since eraseline doesn't work) Jan 19 19:52:12 see the thing is, we are talking about color here Jan 19 19:52:18 which was never part of vt102 Jan 19 19:52:33 so the erasing commands wont mention what to do Jan 19 19:53:33 hmm, i see. do you perchance know what xterm does (is it not 'B'?)? Jan 19 19:53:40 vt102 docs talk about line attributes being reset nto graphics states Jan 19 19:53:51 line attributes are things like double width/height Jan 19 19:54:25 dtzWill, with out having looked at xterms code I can see what it does Jan 19 19:54:29 and i know what wterm needs to do Jan 19 19:56:18 dtzWill, the 3 functions that start on line 330 of terminalstate.cpp Jan 19 19:56:35 see to do just a little more than they currently do Jan 19 19:57:10 TerminalState::erase() prob needs to be overloaded to add a 3rd param Jan 19 19:57:32 currently erase() removes the graphicsstate for the cells it erases Jan 19 19:57:47 yep i see both of that Jan 19 19:57:56 but what needs to happen for the erase line commands is that the graphic state of the erased cells gets set to the the graphics state of the cursor location Jan 19 19:58:00 what's the correct behavior? Jan 19 19:58:06 oh you just said, sorry Jan 19 19:58:38 PuffTheMagic: oh is that it? i thooght you meant they should retai their existing logic Jan 19 19:59:05 so it seems the same as *delete to end of line*, *insert blanks to end of line from cur char* even if we don't do it that way Jan 19 20:00:02 delete to end of line? Jan 19 20:01:01 i guess character attribute does mean graphicsstate Jan 19 20:01:13 err Jan 19 20:01:17 no not exactly Jan 19 20:01:25 Character Attributes: Jan 19 20:01:31 underline, reverse video Jan 19 20:01:33 blink Jan 19 20:01:34 bold Jan 19 20:01:50 so nothing about color there Jan 19 20:02:23 so i bet we could do an even more detailed test Jan 19 20:02:30 1 sec Jan 19 20:02:39 brb meetingness Jan 19 20:10:50 there we go Jan 19 20:10:50 echo -e -n "\e[1J\e[1;20H\e[5;4;1;44mHELLO WORLD\e[6D\e[K\e[m\n" Jan 19 20:11:41 so HELLO WORLD are written in bold, underline and blink Jan 19 20:11:45 with blue background Jan 19 20:11:56 cursor is moved backwards to the space between hello and world Jan 19 20:12:02 and then erase line to end of screen Jan 19 20:12:24 blink,bold,underline are gone Jan 19 20:12:36 but the erased chars have blue background Jan 19 20:13:03 so that behavior does technically meet vt102 specs Jan 19 20:13:17 but i guess terms did what they wanted to for color when that came around Jan 19 20:13:56 OHHHHHHHHHHHHH Jan 19 20:13:58 even better Jan 19 20:14:07 this is true for Erase in Display Jan 19 20:14:22 the command the deletes from cursor to end of display Jan 19 20:14:30 it also sets who screen blue Jan 19 20:14:55 so it seems like we dont need to add a new param to erase() Jan 19 20:15:00 we just need to change is behavior period Jan 19 20:19:57 quite possibly. not having a grasp of what should happen when i wasn't able to say. Jan 19 20:20:06 yay test cases Jan 19 20:26:54 reading that again, it does seem that way. Jan 19 20:35:31 ok i am abandoning work on scrollbuffer Jan 19 20:35:36 as that will be a huge task Jan 19 20:35:41 working on multi cards instead Jan 19 20:41:10 dtzWill, think u have an idea what needs to happen? Jan 19 20:41:19 or at least what I want you to do? Jan 19 20:50:51 multi tabs cards work :D Jan 19 20:52:14 http://ompldr.org/vY2NjZw/wterm_2012-19-01_154943.png Jan 19 20:52:38 looks cool Jan 19 20:53:21 now to figure out how to make them show up in the same stack Jan 19 20:54:00 that is an utter pain to do Jan 19 20:59:07 halfhalo, how do i do it? Jan 19 20:59:43 you have to time it right so that wterm is in the foreground completely before creating a new window Jan 19 21:01:03 so if a user minimised wterm so that they can press wterm icon i take it wterm is not in the fg any more? Jan 19 21:01:20 right. Jan 19 21:01:34 how do I make it in the fg, reactive the window or something? Jan 19 21:01:38 it has to be the full screen active app Jan 19 21:08:09 ugg. I managed to type reboot into my local shell instead of a remote server. Jan 19 21:08:15 hahaha Jan 19 21:08:31 I haven't done that in awhile Jan 19 21:08:44 but PuffTheMagic, when you get the window active event, it still needs a little time before a new window call will stack. Jan 19 21:08:52 like, 20ms padding or something Jan 19 21:08:53 differently colorized shell prompts helps for me Jan 19 21:09:22 my shell was, I just totally didn't see that it had disconnected on me the line above Jan 19 21:11:48 halfhalo, how to i activate the current window with out knowing its name Jan 19 21:12:10 there is an application event that is fired Jan 19 21:13:31 for the name of window? Jan 19 21:13:57 no, when your application is the active app Jan 19 21:14:57 enyo.windows.getWindows()[0] seems like it will do the trick too Jan 19 21:16:17 maybe not Jan 19 21:18:55 if it's an easy change to the mojo version, it could go into wirc-mojo too Jan 19 21:19:21 dwc-, i think halfhalo tried doing that for wirc once Jan 19 21:19:23 * dwc- has his autojoin channels pop up sometimes in a different stack Jan 19 21:19:35 s/a different/its own/ Jan 19 21:19:50 yeah, I had that in it before, but then the code got moved and then removed Jan 19 21:32:00 halfhalo, waiting 100 ms seems to do the trick here Jan 19 21:33:21 dtzWill, just pushed a commit with multi card support Jan 19 21:35:11 there will be edge cases where they still might not stack, but for 99% of the time that will work Jan 19 21:50:43 PuffTheMagic, can you get me a screenshot of wterm and a tagline for wterm? Want to stick it in a little slideshow that will be looping at scale Jan 19 21:56:37 there are screenshots on the github site Jan 19 21:56:42 but I should update them cause they are stale Jan 19 21:57:38 halfhalo, gonna have a demo unit there? Jan 19 21:57:53 yeah, but also http://silentbluesystems.com/ Jan 19 21:58:13 oh god Jan 19 21:58:15 presi? Jan 19 21:58:19 no Jan 19 21:58:21 impress.js Jan 19 22:04:06 it really wasn't all that hard to get going. each slide is just a div. Jan 19 22:05:01 "Oops! Google Chrome could not find silentbluesystems.com" Jan 19 22:05:27 ok, so I'm NOT crazy! my dns DID get messed up! Jan 19 22:06:46 i dont think the two are mutually exclusive :) Jan 19 22:07:59 halfhalo, here Jan 19 22:08:00 https://github.com/PuffTheMagic/wTerm/raw/master/screenshots/htop.png Jan 19 22:08:06 https://github.com/PuffTheMagic/wTerm/raw/master/screenshots/irssi.png Jan 19 22:08:13 https://github.com/PuffTheMagic/wTerm/raw/master/screenshots/midnight_commander.png Jan 19 22:08:20 https://github.com/PuffTheMagic/wTerm/raw/master/screenshots/vim.png Jan 19 22:08:51 halfhalo, do we really need a tag line or can u just say Jan 19 22:09:02 just something to put under the app name Jan 19 22:09:02 wTerm: The webOS Terminal Emulator Jan 19 22:09:09 that wont work? Jan 19 22:09:15 that will work fine Jan 19 22:11:13 halfhalo, if you want to list some features get them from https://github.com/PuffTheMagic/wTerm Jan 19 23:40:10 does anyone know if redirects are handled by the shell or the terminal? Jan 19 23:40:57 what kind of redirects Jan 19 23:41:09 like > Jan 19 23:41:27 that's the shell Jan 19 23:41:48 you don't want your terminal doing io Jan 19 23:49:17 i forgot i need -o with strace Jan 19 23:49:30 dwc-, you in a mood to solve a mystery Jan 19 23:50:04 maybe, shoot. Jan 19 23:50:25 emacs crashes in wterm Jan 19 23:50:29 immediately Jan 19 23:50:33 cant even run help Jan 19 23:50:40 so its someting that happens early in the main Jan 19 23:51:09 I don't have emacs on my touchpad :p Jan 19 23:51:11 i think its related to this: setrlimit (RLIMIT_STACK, &rlim); Jan 19 23:51:18 dwc-, install it then ;) Jan 19 23:51:20 it crashes wterm or crashes emacs? Jan 19 23:51:27 emacs crashes Jan 19 23:51:40 eew, emacs... I didn't get a 32gb touchpad so I"d have room for emacs ;) Jan 19 23:51:59 it could definitely be crashing if it's hitting that rlimit Jan 19 23:52:17 but why would it crash only in wterm Jan 19 23:52:33 is there a var I need to set when I fork the initial term processes? Jan 19 23:52:36 versus sshd or novacom? Jan 19 23:52:54 right Jan 19 23:52:56 or xterm Jan 19 23:53:00 it works with all them Jan 19 23:53:17 check ulimit -a under both Jan 19 23:53:48 there is no -a Jan 19 23:53:53 -s then? Jan 19 23:54:08 hm, doesn't look like you're setting limits in wterm? Jan 19 23:54:27 getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) = 0 Jan 19 23:54:27 kill(5026, SIGABRT) Jan 19 23:54:38 those are the last 2 lines in the strace of emacs Jan 19 23:55:02 lol Jan 19 23:55:12 i think emacs multiples something by the max Jan 19 23:55:31 I have ulimit -a in the shell on the touchpad Jan 19 23:55:37 why don't you... are you using csh or something? Jan 19 23:56:07 is (was) 5026 the pid of the emacs process? Jan 19 23:56:22 i guess it do have a -a Jan 19 23:56:28 and didnt understand its output Jan 19 23:56:53 probably re:5026 Jan 19 23:57:31 hmm Jan 19 23:57:47 where'd you get emacs from? optware? Jan 19 23:58:18 iirc, getrlimit() should be filling in the rlim_cur and rlim_max and ignoring the structure passed in Jan 19 23:58:30 ya optware Jan 19 23:58:43 so it might be checking, finding it's 8192/8192 and not as large as it wants and deciding to die... might need to check the src for that Jan 19 23:58:52 http://opensource.apple.com/source/emacs/emacs-70/emacs/src/emacs.c Jan 19 23:58:57 at least on my touchpad, (using yesterday's build) Jan 19 23:58:59 that does crazy things at startup Jan 19 23:59:08 I have 8192/8192 via sshd as well as on wterm Jan 19 23:59:24 yea you can use -s Jan 19 23:59:32 i had to do login -f root Jan 19 23:59:34 you can get the hard limit by adding -H Jan 19 23:59:36 and then -a worked right Jan 19 23:59:49 weird... I didn't do anything to mine and ulimit -a and ulimit -aH work Jan 19 23:59:53 what's your default shell Jan 20 00:00:14 bash Jan 20 00:00:16 or, do you have any funny things in your shell rc files Jan 20 00:00:39 see what happens if you change it back to /bin/sh (busybox) Jan 20 00:00:46 * dwc- doesn't have bash installed either Jan 20 00:07:45 so if i were to set some limits Jan 20 00:08:21 should I do that in the main of wterm or in the child after the shell process is forked Jan 20 00:11:39 dwc-, u gonna install emacs and help? Jan 20 00:12:07 why are you setting limits? and what are you trying to limit? the plugin itself? or the processes people run in the shell? Jan 20 00:12:24 im trying to stop emacs from crashing in wterm Jan 20 00:12:35 try raising the limits first Jan 20 00:12:35 and i think its because of that infinity shit Jan 20 00:12:41 infinity is good Jan 20 00:12:44 it means unlimited Jan 20 00:13:16 but where should I be setting them Jan 20 00:13:19 in the child process Jan 20 00:13:22 or in the main of wterm Jan 20 00:13:25 for now, do it in testing Jan 20 00:13:34 shell$ ulimit whatever Jan 20 00:13:39 shell$ emacs -nw Jan 20 00:13:45 and see if it starts Jan 20 00:13:55 you're root, right? so you should be able to raise and lower at will Jan 20 00:14:40 both the hard and the soft limits Jan 20 00:15:07 also, I find it odd your "bash" has no ulimit -a ... it's a bash builtin and that is the right bourne-style syntax Jan 20 00:16:33 it does have -a Jan 20 00:16:40 its output it just weird Jan 20 00:16:41 and yes I'm going to install emacs... the sacrifices I make for you. emacs22? Jan 20 00:16:45 and it looked more like a help menu Jan 20 00:16:50 yes emacs22 Jan 20 00:17:58 dwc-: the things we do Jan 20 00:18:09 my poor touchpad has been dirtied Jan 20 00:18:35 dwc-, i know, mine too (this is for rwhitby) Jan 20 00:18:35 it will never be the same Jan 20 00:19:00 well, this is so rwhitby can't say "but it runs in Xecutah fine!" Jan 20 00:19:28 and emacs really is a torture test for vt compatibility Jan 20 00:19:29 well im sure you are not the only emacs user out there Jan 20 00:19:37 tsunami[~]cat .emacs Jan 20 00:19:37 (if (null window-system) (global-set-key "\C-h" 'delete-backward-char)) Jan 20 00:19:37 (kill-emacs) Jan 20 00:19:38 tsunami[~] Jan 20 00:19:47 rwhitby, which is why im glad you brought it up Jan 20 00:20:30 dwc-: not '(exec "vi")' ? Jan 20 00:20:40 dwc-, where is that? Jan 20 00:20:53 my shell host Jan 20 00:21:04 -rw-r--r-- 1 3463 20 85 May 29 2005 .emacs Jan 20 00:21:30 yep, I can duplicate your issue Jan 20 00:21:34 i dont have a ~/.emacs Jan 20 00:21:55 rwhitby: no, but I did play with viper for a couple hours Jan 20 00:22:18 back when I was doing lisp/scheme Jan 20 00:22:37 then near the end of that, I learned about vi's :set showmatch and I never looked back Jan 20 00:22:38 PuffTheMagic: dwc is saying that since 2005 he has had an emacs init script which shuts down emacs if you try and run it. it's a joke. Jan 20 00:23:15 oh Jan 20 00:23:30 the last time I installed emacs on one of my own systems, it was for a gf Jan 20 00:24:25 a gf that runs emacs - she's a keeper Jan 20 00:24:30 no... no she wasn't Jan 20 00:27:01 * dwc- likes them less crazy/insane/needy Jan 20 00:30:38 hmm, that is interesting Jan 20 00:31:27 it runs in old sdlterminal Jan 20 00:31:38 http://silentbluesystems.com/ I call that done unless someone has any changes. Jan 20 00:32:51 Your browser doesn't support the features required by impress.js, so you are presented with a simplified version of this presentation. Jan 20 00:32:55 heh Jan 20 00:33:43 all the links are broken... http://silentbluesystems.com/webos-internals.org Jan 20 00:33:53 PuffTheMagic: it runs in non-plugin mode even D: Jan 20 00:34:04 "The Pre3 was only sold unlocked to consumers." is also somewhat untrue Jan 20 00:34:46 bah, forgot the http and was only officially sold. And that should be a footnote. bleh. Jan 20 00:35:04 Brybry, emacs runs in non plugin mode you are saying? Jan 20 00:35:14 yes Jan 20 00:35:16 with wterm Jan 20 00:35:33 PuffTheMagic: looks like a successful run hits up setrlimit() right after getrlimit() Jan 20 00:35:34 but not in plugin Jan 20 00:35:47 dwc-, yes i know ;) Jan 20 00:36:19 Brybry, probably because as a plugin it gets launched with weird resource limitations Jan 20 00:36:27 I would say, get the source and see what exactly is failing here Jan 20 00:36:42 i linked you yo source ;) Jan 20 00:36:52 you linked to the source that apple uses Jan 20 00:36:57 or, has published at least Jan 20 00:36:57 I'm really sad that HP broke ctrl+c =/ Jan 20 00:37:36 why couldn't they make it do the copy paste nonsense and still pass the event along Jan 20 00:37:40 dwc-, its failing on setrlimit Jan 20 00:42:09 well, it's not an in-general setrlimit failing Jan 20 00:42:53 root@TouchPad:/media/internal# strace ~/setrlimit Jan 20 00:42:57 getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) = 0 Jan 20 00:42:58 setrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) = 0 Jan 20 00:42:58 exit_group(0) = ? Jan 20 00:43:04 runs the same in wterm and via ssh Jan 20 00:43:49 well, except wterm it seems to report two getrlimit() Jan 20 00:44:05 hmm, that is odd Jan 20 00:45:07 dwc-, http://pastebin.com/Ya7txCgU Jan 20 00:47:38 where's LD_LIBRARY_PATH being set? Jan 20 00:47:53 no idea Jan 20 00:48:54 (LD_LIBRARY_PATH is the devil's work) Jan 20 00:49:31 dwc-, are you suggesting this error is a toolchain issue? Jan 20 00:49:40 no Jan 20 00:50:02 I just noticed the linker was trying to pull libgcc_s.so from ... us.ryanhope.wterm though Jan 20 00:50:28 and wondered if you were doing it Jan 20 00:50:33 for some reason or another Jan 20 00:50:37 not on purpose Jan 20 00:52:05 Brybry, does escape work on the bt keyboard? Jan 20 00:53:15 I'll look at this later Jan 20 00:53:23 some interesting things going on here Jan 20 00:53:49 oh hey Jan 20 00:53:50 there's no escape key on the hp bt keyboard Jan 20 00:53:51 nevermind Jan 20 00:53:53 I fixed it. Jan 20 00:53:56 but ctrl+[ doesn't either I don't think Jan 20 00:53:58 "fixed" it Jan 20 00:54:34 Brybry, ctrl+[ doesnt work on the bt keyboard? Jan 20 00:55:15 Brybry, do you know what sym code it sends for that key Jan 20 00:55:21 webos hooks into it as a notification toggle hotkey or something Jan 20 00:55:25 so it probably eats it Jan 20 00:55:32 like it does ctrl+a/c/v now ._. Jan 20 00:55:50 lame Jan 20 00:55:55 I am an unhappy person and will have to figure out a workaround Jan 20 00:56:38 http://ompldr.org/vY2NnZQ Jan 20 00:57:29 dwc-, what did u do? Jan 20 00:58:23 ask ptmalloc3 and libmemcpy what they're doing Jan 20 00:58:45 unset LD_PRELOAD Jan 20 00:58:54 and it'll work Jan 20 00:59:00 not sure why yet Jan 20 00:59:26 not familiar with those two libs, how they're used, or what symbols they export Jan 20 01:00:18 unless someone knows why palm has those loaded, my advice would just be to remove them from the env before you fork the shell Jan 20 01:01:24 wait, hmm ctrl+c/v still work Jan 20 01:01:26 I'm just insane Jan 20 01:01:28 LD_PRELOAD and LD_LIBRARY_PATH Jan 20 01:01:46 I wonder why webos is only eating ctrl+a Jan 20 01:01:57 how do i quit emacs Jan 20 01:02:02 ^X ^C Jan 20 01:02:14 thanks Jan 20 01:02:15 or M-x kill-emacs Jan 20 01:02:16 yeah, I had to read the emacs start to screen to figure that out too =x Jan 20 01:02:36 hit ^G a couple times if ^X^C isn't working Jan 20 01:03:08 (stupid multi-combo things) Jan 20 01:03:19 I'm not sure removing them from the env before forking will actually work Jan 20 01:04:12 actually, I guess it would Jan 20 01:04:34 what about using execvpe Jan 20 01:04:37 instead of execvp Jan 20 01:04:42 and just using a NULL env Jan 20 01:04:46 yea, that should work Jan 20 01:05:04 like /bin/login ignores preserving some environment stuff Jan 20 01:05:05 or including just the things you want to pass Jan 20 01:05:08 or sh resets things or something Jan 20 01:05:30 or at least it did for TERM when I tried to force it Jan 20 01:06:56 thats cause of profile stuff Jan 20 01:07:13 if u want to override TERM create a file in /etc/profile.d Jan 20 01:08:46 yeah, I just put a .profile in /var/home/root instead Jan 20 01:08:57 or that Jan 20 01:09:00 don't have to remount / then Jan 20 01:11:29 what provides execvpe Jan 20 01:11:56 'execvpe' was not declared in this scope Jan 20 01:13:26 oh, maybe there isn't one Jan 20 01:13:36 its a gnu extension Jan 20 01:13:38 u need to define _GNU_SOURCE Jan 20 01:13:43 to enable it Jan 20 01:13:51 but i think i have a header order issue or something Jan 20 01:14:33 can we maybe just not pass -p to /bin/login? Jan 20 01:14:55 ahh Jan 20 01:15:03 ya i can try that Jan 20 01:15:35 looks like that works Jan 20 01:15:52 :D Jan 20 01:16:56 yep Jan 20 01:16:58 thanks Brybry dwc- Jan 20 01:16:59 yesss Jan 20 01:17:05 and that fixes being able to start wterm from within wterm Jan 20 01:17:58 LD_* causes so many problems Jan 20 01:18:45 dwc-, see if u forked wterm u could fix this and get the cred ;) Jan 20 01:19:03 or you can just put it in a commit note or something Jan 20 01:19:20 having the discussion here ended out probably better anyways Jan 20 01:19:32 if I just removed it, brybry might not have suggested removing "-p" Jan 20 01:19:36 =P Jan 20 01:20:21 * PuffTheMagic tags 0.3.3 Jan 20 01:23:33 rwhitby, in about 30 min there will be a new wterm that works with emacs Jan 20 01:23:44 or he can just unset LD_* now :) Jan 20 01:24:29 so now are you going to make a screenshot for emacs Jan 20 01:26:39 sure Jan 20 01:28:06 halfhalo, https://github.com/PuffTheMagic/wTerm/raw/master/screenshots/emacs.png Jan 20 01:28:36 what's so good about emacs Jan 20 01:29:10 egaudet, im not answering that Jan 20 01:29:18 but its another test case for wterm Jan 20 01:29:28 PuffTheMagic: you can't answer that. there is no answer for that question. :p Jan 20 01:30:22 lol Jan 20 01:30:41 when did palm add linux toolchain into sdk Jan 20 01:30:50 recently Jan 20 01:30:54 i was surprised about that too Jan 20 01:30:57 nice Jan 20 01:31:12 i should make an ebuild for the sdk for gentoo Jan 20 01:31:15 now that its complete Jan 20 01:31:50 I might much with gentoo on my new PC one of these days Jan 20 01:31:53 somewhere around 3.0.2 Jan 20 01:35:36 @_PuffTheMagic_ You say that like supporting emacs is a good thing Jan 20 01:35:37 lol Jan 20 02:43:23 Brybry, ping Jan 20 02:44:29 ka6sox-away, i cant reproduce your vi/wterm bug even when im ssh'd in **** ENDING LOGGING AT Fri Jan 20 02:59:57 2012