**** BEGIN LOGGING AT Mon Aug 01 02:59:57 2011 Aug 01 03:13:27 scoutcamper, ping Aug 01 03:28:57 PuffTheMagic, pong Aug 01 04:38:42 scoutcamper, lastest commit added really shitty support for opening files Aug 01 04:38:54 lol Aug 01 04:38:57 ill try it Aug 01 04:50:53 hrm Aug 01 04:51:04 it would help if my TP's battery wasnt dead lol Aug 01 05:06:04 PuffTheMagic, even though i can guess the answer, does the open and save default to /media/internal ? Aug 01 05:06:23 ya i havent added support for traversing dirs yet Aug 01 05:06:30 i needs to make the dialog look less ugly first Aug 01 05:06:49 kk Aug 01 05:07:02 ugly dialog i can live with lol Aug 01 05:59:18 scoutcamper, slightly less ulgy less buggy dialog pushed Aug 01 06:28:55 I need C++ help :-\ Aug 01 06:30:01 that 1 lang i dont touch Aug 01 06:30:20 nor do I, which is why I don't know it well enough to figure out what's going on Aug 01 06:30:34 i like C++, but havent touched it all summer, so my skills are slow Aug 01 06:30:39 and so far google isn't helping Aug 01 06:31:02 C++ is my friend, but ive been learining enyo instead :/ Aug 01 06:31:35 when a derived class is declared thus: derived_constructor_name (parameters) : base_constructor_name (parameters) {...} Aug 01 06:31:48 what order are the constructors called in? Aug 01 06:32:04 oh god Aug 01 06:32:16 i hate constructors Aug 01 06:32:55 but i think its base then derived Aug 01 06:33:31 yeah that would make the most sense I think Aug 01 06:34:27 jacques: are you sure the base is called at all? Aug 01 06:34:43 jacques: or do you have to call super. yourself? Aug 01 06:35:28 scoutcamper, http://ompldr.org/vOXBsNA/tide_2011-01-08_023319.png Aug 01 06:36:34 rwhitby, my reading so far says that the above declaration is when you want both constructors called Aug 01 06:36:44 and doesn't mention having to call it explicitly Aug 01 06:37:06 jacques: in that case, it would have to be base first Aug 01 06:37:36 OK, I'm working on that assumption Aug 01 06:43:55 PuffTheMagic, i like that Aug 01 06:54:47 i got to get them sorted Aug 01 06:54:54 then i will add traversing dirs Aug 01 06:54:59 tomorrow Aug 01 07:04:18 PuffTheMagic: lmk when you use TIDE to work on TIDE Aug 01 07:04:23 self-hosting editor or something Aug 01 07:04:26 ;) Aug 01 07:05:00 preware needs dropbox support Aug 01 07:05:13 so that u can just install an ipk like its a url Aug 01 07:05:40 even if we could, no compiling on TP makes working on TIDE with TIDE a pain Aug 01 07:05:59 why can't you compile on-device? :) Aug 01 07:06:00 why no compiling? Aug 01 07:06:12 through chroot yes Aug 01 07:06:20 pslm-package Aug 01 07:06:28 palm-package, no* Aug 01 07:06:39 ipkg-opt has a good part of it, idk about w.r.t that node/npm craziness :) Aug 01 07:06:47 craziness only because i haven't used it, to be clear :). Aug 01 07:06:51 lol Aug 01 07:06:55 I DONT LIKE WHAT I DONT UNDERSTAND Aug 01 07:06:56 yes!!! Aug 01 07:06:57 xD Aug 01 07:07:19 I finally got this thing displaying other than a black screen Aug 01 07:07:27 lol Aug 01 07:07:28 jacques: what're you working on? Aug 01 07:07:31 fun with shaders? Aug 01 07:07:39 gah i'm jealous, i miss porting over new exciting shit Aug 01 07:07:46 these days i just maintain things :( Aug 01 07:08:42 for some definitino of "fun" Aug 01 07:08:51 nah, it's a blast Aug 01 07:09:10 and you getta be all excited that you even got it working, never mind no controls or usability :P Aug 01 07:09:11 lol Aug 01 07:09:13 I'm trying to get a bit of understanding of gles2 Aug 01 07:09:34 because that's where the performance is on adreno220 Aug 01 07:09:38 (no hardware 1.1) Aug 01 07:09:52 * dtzWill nods Aug 01 07:10:28 PuffTheMagic: is there some open source dropbox webOS code we can use? Aug 01 07:10:58 (can you not point curl at a dropbox url?) Aug 01 07:11:13 or are we talking preware having listing of files, etc? Aug 01 07:11:34 yes, you can drop a URL in the main screen of Preware if you like Aug 01 07:11:42 I suspect PuffTheMagic is talking about dropbox browsing support Aug 01 07:14:17 heh, looks like I need to tirn off vsync Aug 01 07:24:13 jacques: did you figure out how to turn off vsync? lmk if you do Aug 01 07:24:28 not saying it's impossible; i don't remember--but knowing for sure (and how) would be useful/nice. Aug 01 07:25:10 dtzWill, I am just realizing that it's going to be harder than I thought, as I already have SDL_GL_SetAttribute(SDL_GL_SWAP_CONTROL, 0); Aug 01 07:25:25 jacques: yeah... eglSwapInterval? Aug 01 07:25:26 which is having no effect Aug 01 07:25:50 lo Aug 01 07:25:52 lol :( Aug 01 07:25:53 dtzWill, hmm, I'll have to look into that one Aug 01 07:26:18 SDL_GL_SWAP_CONTROL is having no effect, I haven't tried eglSwapInterval Aug 01 07:27:15 hmm, SDL_GL_SWAP_CONTORL _should_ work i'd think? Aug 01 07:28:01 yep, the sdl sources have it propagating to gl_config.swap_control, and elsewhere check if that's not -1 then /they/ call eglSwapInterval Aug 01 07:28:23 when the gl context is made current, so that should work :( Aug 01 07:28:39 (in a 30-second bounce through the sdl source lol) Aug 01 07:28:47 (salt accordingly) Aug 01 07:29:02 dtzWill, I am calling it before SDL_SetVideoMode() but I just read something which said it should be called after Aug 01 07:29:07 really do wish we could build sdl ourselves... Aug 01 07:29:16 dtzWill, ++++++ Aug 01 07:29:19 jacques: yes, afterwards Aug 01 07:29:29 dtzWill, and 0 is off ? Aug 01 07:30:01 not sure. -1 is disabled, i would imagine 0 is off yes Aug 01 07:31:07 swap_control gets set to -1 in SDL_VideoInit, which is called when you do your SDL_Init( .. SDL_INIT_VIDEO ...) Aug 01 07:31:19 regardless, you should be able to SDL_GetAttribute on it to see? Aug 01 07:32:07 fwiw SetVideo calls SDL_Init for you automagically if you're lazy :), just saw that in the source hehe Aug 01 07:32:09 hmm, moving it after the SDL_SetVideoMode() made no difference Aug 01 07:33:15 before SetvideoMode, after SDL_Init would be my guess, but regardless either you're setting it and it sticks, or not.... Aug 01 07:33:25 maybe jsut make the call to eglSwapInterval yourself and be done :P Aug 01 07:33:32 ...if you have the egl display or w/e >_> Aug 01 07:33:53 but SDL_SetVideo calls the MakeCurrent that issues the eglSwapInterval iff swap_control != -1 Aug 01 07:36:35 hmm, I'll move it back between SDL_Init and SDL_SetVideoMode and read some more Aug 01 07:37:47 lolol Aug 01 07:37:50 fbprogress -f erasing Aug 01 07:38:05 hahaha Aug 01 07:38:19 just kill fbprogress and tap screen (anything to cause screen update) and all is well Aug 01 07:38:21 haha Aug 01 07:38:25 delicious Aug 01 07:41:20 fbprogress? Aug 01 07:41:56 oh, I see Aug 01 07:52:37 lol :) Aug 01 08:31:08 if i made the xserver's rendering a good deal faster but lost ability to change orientations, you think that'd still be useful? Aug 01 08:31:28 rotations would still work, but rotating in software is turrible :( Aug 01 08:31:48 which orientation would be fast? landscape? Aug 01 08:34:54 jacques: ye Aug 01 08:34:56 *yep Aug 01 08:34:59 right home button on right Aug 01 08:36:53 would probably be very useful for ppl using ts-only apps or with bt keyboards Aug 01 08:37:08 well i'm actually not sure of default framebuffer orientation on the phones Aug 01 08:37:18 nor al I Aug 01 08:37:20 m Aug 01 08:37:58 default (and only) resolution is 320x480 Aug 01 08:38:07 so actually i think this'd work fine on phones too Aug 01 08:44:54 haven't gotten around to testing on phone yet Aug 01 08:46:49 dtzWill: hmm - rotation is a pretty significant feature Aug 01 08:47:53 rwhitby: i don't know that i see why, but am open to convincing :). Aug 01 08:48:05 but even if it is, it doesn't seem like an easy decision Aug 01 08:48:38 if i was superman i could do something slick and use GLES for other orientations (but slower due to texture uploading), but raw framebuffer for default orientation Aug 01 08:48:50 but that's ...crazytalk Aug 01 08:48:52 lol :P Aug 01 08:53:06 dtzWill: I'm always wary of a decision which permanently removes existing functionality Aug 01 08:53:13 yeah there's not nearly enough GPU memory to do rotation in the GPU Aug 01 08:53:35 dtzWill: especially when previously we were wondering about how to do on-the-fly rotation Aug 01 08:55:32 rwhitby: well fwiw on-the-fly rotation is relatively easy to support Aug 01 08:56:17 takes a bit of attention to detail, but nothing fundamentally difficult about it. have it mostl working with framebuffer atm, just by enabling randr Aug 01 08:56:28 add code to check accel periodically or whatever we prefer and do that Aug 01 08:56:32 but yeah, hmm.... Aug 01 12:17:22 o______O has the pdk documentation always said things like "This section and PDL calls do not apply to desktop development"?? Aug 01 14:41:05 how do you handle the gesture input for SDL running on 1.4.5? I get a translatekey error on 1.4.5 only. it seems like SDLKey 231 (SDLK_World_71) does not exist on the 1.4.5 SDL. anyone run into this? Aug 01 14:49:19 or is there a way to see what sdlkey is being sent? Aug 01 14:54:57 elpollodiablo1: just log the value :) Aug 01 14:55:17 and prety sure 1.4.5 _does_ support all kinds of gesture events, i use them in misc apps that def worked on 1.4.5 Aug 01 14:56:27 on 1.4.5 back gesture is 'ESC' iirc Aug 01 14:58:36 back is 27, forward is 229, gesture button is 231 Aug 01 14:58:49 are you saying they don't #define it in their headers? Aug 01 14:59:08 i have no idea about that, just define it neatly in your own code and go from there. Aug 01 15:17:20 dtzWill: the weird thing is I have gestures causing no issues in another one of my programs, but for some reason any gesture input segfults that app and say the TranslateKey() failed in libsdl Aug 01 15:18:33 did you initilize the PDL stuff or just straight SDL? Aug 01 15:21:46 elpollodiablo1: in 1.4..5? probably just straight sdl Aug 01 15:21:53 also i don't think i've used translatekey on it Aug 01 15:21:58 so maybe that breaks, idk Aug 01 15:22:39 alright thanks, just trying to see if anyone else ran in to this. Aug 01 15:42:47 rwhitby: what about dropbox, i dont remember what I said? Aug 01 15:43:14 did it have to do installing ipks from dropbox? Aug 01 15:43:52 i was imagining just using the synergy provider apis like the other TP apps do Aug 01 15:43:56 but they both have open apis Aug 01 15:44:10 so it would't be hard to make our own calls to the service Aug 01 15:52:33 halfhalo-work: yo Aug 01 15:52:41 hmmm? Aug 01 15:53:03 halfhalo-work: i fixed that auto caps issue, its all working now Aug 01 15:53:11 sweet Aug 01 15:53:35 halfhalo-work: and the code is in a repo now if you want to test Aug 01 15:53:46 sweet Aug 01 15:53:59 enyo/editah Aug 01 15:54:16 clone then run 'make prepare' to fetch the submodules and auto run npm Aug 01 15:56:58 or not. doesn't like to grab submodules without a key. Aug 01 15:58:54 oh ya Aug 01 15:58:54 that 1 submodule :/ Aug 01 15:59:00 i need to figure out how to get it to pull from git:// but push to git@ Aug 01 15:59:35 does the submodule have a package.json file in it? Aug 01 16:00:29 idk Aug 01 16:00:29 i dont acually use npm for them Aug 01 16:02:03 npm is only to pull in uglify-js Aug 01 16:02:04 err Aug 01 16:02:04 maybe thats not right Aug 01 16:02:04 it does "build" some stuff Aug 01 16:02:04 im still a node noob Aug 01 16:02:33 well, you can, and you can make it so that instead of git submodules they get setup properly for end users, while still working for development as well. and npm's modules stuff is a great thing. lets you not have to deal with npm on other systems Aug 01 16:03:09 well if you know how to convert the setup let me know Aug 01 16:03:22 or do it and give me patches Aug 01 16:28:19 PuffTheMagic: don't see clean way to do it, but git remote set-url --push origin git@git.webos-internals.org:js/ace Aug 01 16:28:28 does what you want to set the push url independant of the pull URL Aug 01 16:28:48 (sets 'pushurl' in .git/config for that remote) Aug 01 16:29:13 unfortunately adding similar pushurl entry in .gitmodulies doesn't seem to work, so you'd have to modify the remote after cloning if you were going to do that. shrug. Aug 01 16:29:24 :/ Aug 01 16:29:30 i guess i could just add that to make prepware Aug 01 16:29:33 prepare Aug 01 16:30:10 yeah, it's not the best, i wouldn't judge you badly if you didn't >_> lol, up to you Aug 01 16:30:19 but was curious, looked into it, and thought i'd share the results :) Aug 01 16:30:41 thanks Aug 01 17:14:42 dtzWill: my connection sucks right now can you do a fresh clone of enyo/editah then make prepare and see if eveything is ok Aug 01 17:15:06 PuffTheMagic: sure thing Aug 01 17:15:23 PuffTheMagic: do you care if i checkout editah itself as git:// or git@ ? Aug 01 17:16:10 dtzWill, i did git@ and it worked, try git:// Aug 01 17:17:03 PuffTheMagic: git@ works here too :) Aug 01 17:17:45 the only difference is that one URL will have push and pull, the other only pull Aug 01 17:17:48 PuffTheMagic: git:// works here too, but shouldn't be Aug 01 17:18:13 PuffTheMagic: lol commit log message Aug 01 17:18:15 "updates" ;) Aug 01 17:19:01 PuffTheMagic: errr Aug 01 17:19:11 PuffTheMagic: you have the url's a level off? Aug 01 17:19:12 git remote -v Aug 01 17:19:17 origin git://webosinternals.org/js/ace (fetch) Aug 01 17:19:17 origin git@git.webos-internals.org:js/pilot (push) Aug 01 17:19:27 origin git@git.webos-internals.org:enyo/editah.git (fetch) Aug 01 17:19:27 origin git@git.webos-internals.org:js/ace (push) Aug 01 17:20:00 want 'cd ace' before first set-url command, 'cd ace/support/pilot' (or whatever) before second one, probably Aug 01 17:20:49 opps Aug 02 00:32:08 Simon80: 'ello :) Aug 02 00:32:36 hi :) Aug 02 00:32:45 this isn't private or anything, but it's dev-talk _only_, which is useful to both stay focused on dev when dev'ing and keep user discussion visible in its own right Aug 02 00:32:51 yeah Aug 02 00:33:03 oh, you can read? Aug 02 00:33:12 * dtzWill looks to topic and back to Simon80 and sighs Aug 02 00:33:14 :) Aug 02 00:33:19 lol Aug 02 00:33:21 sorry :) Aug 02 00:33:30 no, I didn't mean it that way Aug 02 00:34:13 Simon80: so i'm gonna take a whack at removing binaries entirely from the repo Aug 02 00:34:18 that's nice Aug 02 00:34:22 i suppose i could merge in your changes and leave it at that.... hrm. Aug 02 00:34:30 yeah, I was going to say you should start with that Aug 02 00:34:33 just cherry pick it Aug 02 00:34:33 shame that everything's in that src subtree Aug 02 00:34:49 that's fixable too Aug 02 00:34:50 can we move and keep history neatly? hrm Aug 02 00:35:19 all you have to do is make the move with as few other changes as possible, and git merge can usually figure it out Aug 02 00:35:44 Simon80: reason to not merge your master branch over, instead of cherry-picking misc commits? Aug 02 00:35:48 I was actually considering transplanting your whole history onto upstream's, so I could try to upstream things Aug 02 00:35:55 oh, no reason Aug 02 00:36:04 you'll have conflicts with your latest stuff though Aug 02 00:36:04 Simon80: upstream? Aug 02 00:36:08 VBA upstream Aug 02 00:36:29 there...is an upstream these days? Aug 02 00:36:34 I guess not Aug 02 00:36:38 I think they're still sort of alive Aug 02 00:36:53 I have to try to upstream at least some of this stuff Aug 02 00:36:53 fwiw i pulled 1.8.0 in and got it working, but it was always slower which was :( Aug 02 00:36:59 ah, yeah Aug 02 00:37:03 haha mkay, well i'm for that Aug 02 00:37:09 but like, VBA is so broken in some ways Aug 02 00:37:12 i'd imagine someone hosts patched versions for debian or something Aug 02 00:37:17 it could be really low hanging fruit to fix Aug 02 00:38:06 like, my AFS changes are objectively better, I think, then automatically twiddling with a fixed frameskip know Aug 02 00:38:07 knob Aug 02 00:38:55 I guess what I'm saying is there's a lot of room for VBA to be better Aug 02 00:39:25 the Debian packages aren't even usable right now... I guess I have to wonder what upstream is doing then Aug 02 00:39:28 if 1.8 is slower Aug 02 00:39:30 etc. Aug 02 00:39:45 but they're still the rally point where other porters get code from Aug 02 00:39:50 fragmentation is bad Aug 02 00:40:16 well it's probably one of those "we cleaned up the code, and made it work better, but lost the performance from all our manual inline macro-tastic goodness" Aug 02 00:40:31 maybe Aug 02 00:40:48 I like to think compilers aren't that bad, but it depends what the cleanup entails, I guess Aug 02 00:40:52 i don't think vba did the same thing snes9x did, which is decide to stop focusing on speed and focus on accuracy since the host machines could handle it properly Aug 02 00:41:05 Simon80: I similarly like to have faith in my compiler :) Aug 02 00:41:36 yeah, and that's great, but it doesn't excuse gratuitous performance problems, either Aug 02 00:42:12 w/e though Aug 02 00:42:56 I have to give credit to them for having a working emulator, that's more important Aug 02 00:43:13 anyway, I've lost my point Aug 02 00:46:53 I was wondering if you could give me any links to find out more about your Xorg related blitting hack Aug 02 00:47:31 I might be finding it, actually Aug 02 00:47:56 oh, nvm Aug 02 00:48:21 Simon80: oh it's stupid, lol. Aug 02 00:48:39 the only insight was that luna sets up /dev/fb1 for the card, and if you scribble to it directly it can be much faster Aug 02 00:48:44 while still letting you be inside a card Aug 02 00:48:55 as opposed ot scribbling on /dev/fb0 just writes on top of what luna presents Aug 02 00:49:23 so just initialize sdl with GL, and then ditch that and scribble to /dev/fb1 yourself Aug 02 00:49:40 yeah, so I'm wondering how I do that Aug 02 00:50:06 I guess I'm supposed to know about those fb devices Aug 02 00:50:11 but I know nothing, lol Aug 02 00:50:13 :P Aug 02 00:50:58 Simon80: essentially you mmap the framebuffer into your address space, and make ioctl's to the device to tell it to update itself Aug 02 00:51:08 i'm rather new to it as well, but that was enough to get me going for what i needed Aug 02 00:51:23 got half-way through implementing it when remembered the fbdev xorg stuff and merged most of that in :) Aug 02 00:51:45 oh, so you didn't have to write that part because you could just config X to do it Aug 02 00:52:19 (C) 1999, lol Aug 02 00:53:09 indeed, although if you take a look at the sdl sources you can see how palm does it, which is what gave me the idea Aug 02 00:53:26 SDL on the Pixi doesn't work though Aug 02 00:53:35 hmm, but it probably works on the Veer Aug 02 00:53:54 the merge mostly got me things like querying it for modes (which doesn't work since it only reports the one mode) and randr 'support' by advertising rotations and setting the appropriate software-rotation algorithm Aug 02 00:54:02 but yeah it was rratthherrr useful :) Aug 02 00:54:18 Simon80: i'd imagine so. not sure what the exact issue with SDL on pixi was, other than "didn't work" or something. Aug 02 00:54:43 you get a frame out (maybe a couple? I don't know how it breaks), and then it stops Aug 02 00:54:46 stops updating Aug 02 00:55:05 if I recall.. it might have just been a black screen for me, it's been a couple of months Aug 02 00:55:26 but the metaviewsoft guy describes it as working for a short time Aug 02 00:57:23 I'm going to see what changed in SDL between WebOS versions Aug 02 00:58:36 interesting. Aug 02 00:58:47 oh, nvm, apparently there's no PDK stuff at all on the Veer as of May? Aug 02 01:03:25 oh, sweet, I have a post from unwiredben on May 5th saying they fixed the Pixi's SDL problem during the 2.0 cycle Aug 02 01:04:59 oh, nice! :D Aug 02 01:06:07 Simon80: having fun with merges, will let you know when things seem okay :) Aug 02 01:06:11 oh, yeah Aug 02 01:06:17 use git mergetool Aug 02 01:06:32 might _slightly_ break things for you, i'll try to keep things reasonable, and am happy to discuss compromises where needed Aug 02 01:06:36 oh i did the actual merge itself Aug 02 01:06:40 oh, ok Aug 02 01:07:06 yeah, messing with the build part is unpleasant to me too Aug 02 01:07:53 I think most of the settings could go as configure options like --webos-device=pre or something Aug 02 01:07:59 but I'm too lazy to do it Aug 02 01:08:06 at the moment at least Aug 02 01:10:54 yeah taht's the kinda of thing i'm working at Aug 02 01:11:01 sorry :( Aug 02 01:11:16 try to make your commits as small as possible, btw Aug 02 01:11:20 easier to merge them Aug 02 01:11:24 git add -p Aug 02 01:11:28 I don't mind at all, just if I make a decision that seems wrong, rubs you the wrong way, or makes your dev cycle painful lmk and well fix things Aug 02 01:11:35 sure Aug 02 01:11:47 I think it will be fine, since we're building the same way now Aug 02 01:12:03 ah, i use git add -i and then migrate down to patching, but same idea. regardless i'll try to keep patches smaller :) Aug 02 01:12:07 I kind of went around your config script before because it was passing -L/usr/local/lib and stuff Aug 02 01:12:30 I didn't know about sb2, so it was confusing Aug 02 01:12:34 it all makes sense now Aug 02 01:12:49 yeah and then i removed the sb2 stuff and some of that lingered, sorry Aug 02 01:12:55 np Aug 02 01:13:11 working by myself, sometimes discipline regarding those kinds of things slips :3 Aug 02 01:13:39 don't worry about it, I'm just kind of picky about that stuff I guess Aug 02 01:14:57 working build system, how unreasonable ^.^ Aug 02 01:15:03 lol Aug 02 01:15:23 no, I mean, small commits, and having build scripts that don't force a particular set of paths on the developer Aug 02 01:15:39 I'm really meticulous about commits though Aug 02 01:16:18 I figure it's easier to review, revert, and merge them if they're small Aug 02 01:16:19 scoutcamper, ping Aug 02 01:16:48 and when you're merging you can fix conflicts more easily if you know exactly what the conflicting commit was supposed to change Aug 02 01:16:57 * dtzWill nods Aug 02 01:17:12 I might take it too far though Aug 02 01:17:21 I certainly could learn to manage my workflow better esp w.r.t git :), and regardless i'm happy to meet halfway Aug 02 01:17:32 and since you're the one doing the active dev these days, hardly my place to be unreasonable hehe Aug 02 01:17:38 heh Aug 02 01:18:05 I can't own the thing though, I only have a Pixi Aug 02 01:18:11 PuffTheMagic, pong Aug 02 01:18:16 * dtzWill nods Aug 02 01:18:33 and I feel like I don't deserve that much credit so far Aug 02 01:18:54 scoutcamper, you ready to pull in a few min Aug 02 01:19:05 massive enhancements to file dialog Aug 02 01:19:37 I mean, I definitely improved the perf by a lot, but compared to the work you've done Aug 02 01:19:54 and the whole wosi accomplishments Aug 02 01:20:05 PuffTheMagic, yep, ready to pull and build and run :D Aug 02 01:20:26 Simon80: so how do you handle the toolchain/libstdc++ issue atm? Aug 02 01:20:39 I link dynamically, that's why I use such an old version Aug 02 01:20:49 I intend to switch to 2011.03 as well though Aug 02 01:20:53 so don't worry about that Aug 02 01:20:56 I see you're buliding using stock toolchain. I just replaced /opt/PalmPDK/arm-gcc to point to a newer toolchain and go from there Aug 02 01:21:05 but not sure it's reasonable to expect others to have that setup Aug 02 01:21:18 the stock one will work fine with static libstdc++ Aug 02 01:21:22 there's just no reason Aug 02 01:21:30 also, I'm on Linux, there's no stock toolchain ;) Aug 02 01:21:37 but 1.4.5 uses arm2007q3 Aug 02 01:21:51 if you go newer, the binary needs a newer libstdc++ Aug 02 01:22:12 no reason to have static libstdc++ Aug 02 01:22:30 i'm sorry, i think we're miscommunicating. Aug 02 01:22:42 i'm not arguing, just don't understand Aug 02 01:22:54 it was my undestanding 2007q3 could not compile vba at all Aug 02 01:22:55 sorry, I'm kind of disjoint Aug 02 01:22:58 oh, no, it can Aug 02 01:23:08 hooray there's our issue. Aug 02 01:23:12 it just uses up a huge amount of memory during GBA.cpp Aug 02 01:23:22 i was using a newer version b/c iirc it did something wrong Aug 02 01:23:22 like, 1.5GB or something if I recall Aug 02 01:23:27 it's pretty wrong Aug 02 01:23:27 oh shit am i confusing this with snes9x Aug 02 01:23:29 i might be. Aug 02 01:23:43 nah i got the memory to spare, i thought it flat out produced incorrect code/failed to assemble. Aug 02 01:24:23 yeeeep i confused this with snes9x. sigh. Aug 02 01:24:58 okay so only reason to use libstdc++ is to leverage newer compilers, and iirc upgrading from 2007q3 to...2010q1 or w/e at the time was marginally faster Aug 02 01:25:23 I think the memory issue may have been a reason too Aug 02 01:25:32 it sometimes errors out for lack thereof Aug 02 01:25:51 nah, i care less about that. if it's an issue can't you tell gcc to increase i'ts memory limit? Aug 02 01:26:08 i might be confusing it with misc llvm tools that have that option... Aug 02 01:26:29 stock toolchain meant the one palm used, and ships with their PDK for the target platform :) Aug 02 01:26:35 I know Aug 02 01:26:45 there's no Linux PDK though Aug 02 01:26:54 who needs that Aug 02 01:26:56 uhhh okay so you're suggesting ship a copy of libstdc++ with our app if we use a newer toolchain? Aug 02 01:26:56 ;) Aug 02 01:27:01 no, no Aug 02 01:27:05 :( Aug 02 01:27:05 bad idea Aug 02 01:27:09 you do what you did Aug 02 01:27:18 or pass --static-libstdc++ or whatever it is Aug 02 01:27:24 added in 4.5 Aug 02 01:27:47 oh they have an option for that now? hooray. was using -print-file libstdc++.a lol >.< Aug 02 01:28:14 it's pretty recent though Aug 02 01:28:22 I think that's part of why I didn't try it Aug 02 01:28:34 http://gcc.gnu.org/gcc-4.5/changes.html Aug 02 01:29:53 I can't see it in the help of 2010.09 though Aug 02 01:29:56 try and see Aug 02 01:30:43 scoutcamper, ok Aug 02 01:30:44 there's also gcc -dumpversion if you want to be flexible Aug 02 01:35:42 PuffTheMagic, i like the way it looks, but, it isnt listing the contents of the directorys Aug 02 01:36:11 how are you packaging t Aug 02 01:36:15 make Aug 02 01:36:32 is it packaging the service correctly? Aug 02 01:36:45 seems to be Aug 02 01:37:00 do you have 10000000 files in your /media/internal dir? Aug 02 01:37:09 no Aug 02 01:37:20 you getting any errors Aug 02 01:37:22 it works here Aug 02 01:37:31 let me boot lumberjack Aug 02 01:39:25 tis working here :/ Aug 02 01:39:38 do you have the latest sdk? Aug 02 01:39:48 yes Aug 02 01:40:16 but lumberjack isnt cooperating Aug 02 01:41:32 :( Aug 02 01:41:43 PuffTheMagic, after device reboot, tide works Aug 02 01:41:52 as does lumberjack Aug 02 01:44:50 \ Aug 02 01:45:13 scoutcamper, latest commits sorts the entries as one would expect Aug 02 01:45:28 oil, got a bad ass file picker brewing for enyo ;) Aug 02 01:46:12 compiling Aug 02 01:48:06 PuffTheMagic: the new filepicker is neat (in 3.0.2), it now grabs synergy photos! Aug 02 01:48:49 that has more to do with the indexer Aug 02 01:49:04 palms file pickers are always "meh" Aug 02 01:49:11 ++ Aug 02 01:49:15 flat list of files :/ Aug 02 01:49:42 in 3.0 it only showed the internal tp pictures, in 3.0.2 it now shows all the albums that are displayed in the photo app Aug 02 01:50:11 PuffTheMagic, it takes about 10 sec to populate folders and stuff, but it does work Aug 02 01:50:12 big deal Aug 02 01:50:16 ^^ NuttyBunny Aug 02 01:50:28 scoutcamper, ony big dirs seem slow Aug 02 01:50:37 like internal Aug 02 01:50:40 rest seem fast Aug 02 01:50:43 im sure it can be improved Aug 02 01:51:04 PuffTheMagic, yeah, OC's would do that ;-) Aug 02 01:51:07 PuffTheMagic: it is to me ;) having the user download his picture first and then pull it to the app was not acceptable Aug 02 01:51:38 scoutcamper, OC isnt a fix Aug 02 01:52:04 NuttyBunny, again, that has more to do with the indexer than the actual file picker widget Aug 02 01:53:23 ok, I won't argue. 3.0.2 is here and works for me :) Aug 02 01:55:58 oil, http://ompldr.org/vOXB6Yw/tide_2011-01-08_215401.png Aug 02 01:56:02 NuttyBunny, ^^ Aug 02 01:56:51 weird split view :/ Aug 02 01:57:07 better then palms though :) Aug 02 01:57:55 if it wasnt split then it would be 1 massive long list still Aug 02 01:58:26 i guess I could go back to single view Aug 02 02:37:53 Simon80: fwiw, between AFS updates and the sound stuff, GBA stuff on the pre is much better Aug 02 02:38:28 at least when i'm testing super mario, the sound is great and with your cahnges the playback is smooth, as opposed to it being clearly visually lamer in the app cat version atm Aug 02 02:38:41 also, misc testing says your changes work fine on the pre as well Aug 02 02:38:45 :) Aug 02 02:41:06 I think underflow might be handled better if I add back a wait condition in the sound callback function (to wait for sound to be available if it isn't yet, instead of stuttering), but it always feels kind of wrong to consider that Aug 02 02:41:46 it wouldn't reintroduce any deadlock bugs though, and it's not really that bad, it's not pro audio, just an emulator Aug 02 02:42:14 I'm really glad those changes helped on the Pre as well though Aug 02 02:42:23 you should have seen GB on the Pixi before that Aug 02 02:42:33 I mean, you pretty much did **** ENDING LOGGING AT Tue Aug 02 02:59:57 2011