**** BEGIN LOGGING AT Mon May 6 00:25:45 2002 --> stile (~adiamas@as5300-9.216-194-23-71.nyc.ny.metconnect.net) has joined #rockbox --- Topic for #rockbox is Open Source Jukebox Firmware - http://bjorn.haxx.se/rockbox/ --- Topic for #rockbox set by Zagor at Fri Apr 12 09:45:52 <-- adiamas has quit (Killed (NickServ (Nickname Enforcement))) --> lou_soyur (~adiamas@pool-151-204-140-140.ny325.east.verizon.net) has joined #rockbox --- You are now known as adiamas --> [TDM]Mr_B|away (~nx_2000@ip68-0-62-70.no.no.cox.net) has joined #rockbox --- [TDM]Mr_B|away is now known as [TDM]Mr_B --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox --- ChanServ gives channel operator status to Bagder <[TDM]Mr_B> hello? hey ho <[TDM]Mr_B> ...just an outsider with some questions go ahead and ask, I'll try to answer <[TDM]Mr_B> i've been following this project for a short while and i'm very interested... <[TDM]Mr_B> I was wondering what progress you've made. Is there anything working that a programming novice can use? we're doing two different simulators, for running the app itself on host machines one for win32, one for X11 <[TDM]Mr_B> is there an exe of that sim? we do not have the app actually running on target just yet edx is the author of the win32 simulator, we should ask him to post a binary on the web site <[TDM]Mr_B> exactly what improvments/additions are you guys working on? I think many of them are mentioned of the web site we need to implement everything by ourselves so we will try to write an mp3 player software that meets the audience demands <[TDM]Mr_B> when do you think some of these enhancments will be avaliable for the layman to download? we *hope* we will have the first actual target firmware that works *somewhat* before the end of this month <[TDM]Mr_B> cool <[TDM]Mr_B> how well does tretris work? <[TDM]Mr_B> ...tetris it works pretty good ;-) no scores yet though I expect that we will get even more programmers joining up the very second we have a first target version <[TDM]Mr_B> i'm looking forward to all of it :-) me too <[TDM]Mr_B> I see you guys are working on a FAT 32 fs yes, we need that to be able to read the disk <[TDM]Mr_B> I guess that would make installation/use easier? the disk is using fat32, we need to read dirs/files from it... <[TDM]Mr_B> ok, I get it <[TDM]Mr_B> would you mind tellingme exactly what other modifications you guys are working on? we replace the entire software we hope to make the player more intuitive and more capable than the built-in software if you've used it, you know it isn't very good its about resuming play in a playlist handling big playlists and all sorts of stuff <[TDM]Mr_B> well, thanks for entertaining my questions personally, I do most of this because of the shear fun <[TDM]Mr_B> same here, only with coding hypertext * adiamas coughs "Read the faq" feel free to let us know if you think you can help out the project in any way you see fit * adiamas nods hey Bagder whats doing? did lots yesterdat, as the CVS-list shows ;-) heheh yeah.. I _can't_ believe that i forgot to call the function pointers when i put that in :) hehe <[TDM]Mr_B> ...I read the FAQ 3 times heheh k :) can't help; it... its my pet project :) any suggestions on additions or rewording? adiamas: "why do you write a replacement software?" adiamas: "when do you think there is a version available for my Archos?" <[TDM]Mr_B> well, it doesn't answer too many questions <[TDM]Mr_B> tell people about exactly what's being redone and maybe when people will be able to enjoy your labors on their players What exactly are you trying to achieve with this line of development? should that be reworded? adiamas: could you unmber the questions? it would be easier to direct people to a specific question then hm number hmmm how shoul di answer "when do you think there will be a version available?" sure could... adiamas: well, I think that question deserves to be answered very unspecific we might all get very busy all of a sudden so say something like maybe in a month, maybe in three, maybe in six <[TDM]Mr_B> true <[TDM]Mr_B> to quote Dilbert... "ONe month to complete the project and 5 months to play 'Doom' on my computer." lol in this case, of course, it is Tetris ;-) * Bagder just applied patched mailed to him from Dave Chapman patches grrrr, can't type hmmm ... i need a polite way of saying 'Fuck off you complete moron." Suggestions? heh, in which context? never mind.. i have an answer... you're the american ;-) Mr_B: where are you from? --> calpefrosch (~calpefros@62.52.174.30) has joined #Rockbox moin morning calpefrosch the win32 simulator might need some attention now ;-) * Bagder whistles innocently hmm, let's talk about some details. * Bagder browses subdirs in his simulator! calpefrosch: what kind of details? <[TDM]Mr_B> ...New Orleans okay.. faq updated.. anyone wanna take a look? * Bagder shows his thumbs up :) im a bit heavy on the sarcasm.. but bugger it... let them deal with it... we need that <[TDM]Mr_B> ...an improvment...and I love sarcasm it makes it more obvious everything isn't so serious * Bagder runs for more coffee * adiamas cheers wildly [TDM]Mr_B your on windows correct? <[TDM]Mr_B> Windows XP home <[TDM]Mr_B> well, I should reall y be in bed, night folks <-- [TDM]Mr_B has quit () --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox --- Zagor_ is now known as Zagor morning Zagor morn you sure had a busy sunday :) yeps, did my share of cleaning up ;-) sorry, have been to a meeting Zagor: the simulator is growing cooler nice! Dave Chapman stepped in and sent me patches too --> Linus (~linus@labb.contactor.se) has joined #rockbox Boo! * Bagder boos Linus back looking at the throughput that this guy dave is doing.. maybe he should be considered for cvs access... yes, I mailed him about it cool Badger: I ment which kind of attention ? calpefrosch: ah, it won't build just now since I've added more firmware files to the build * adiamas screams * adiamas slams his head against the wall.. keeping both simulators up-to-date is a challange okay.. i screwed up.. so i may as well clue everyone in... * adiamas was working on tic-tac-toe for the archos... * adiamas was 1/2 way there... adiamas is a hero Badger: ok, I'll try to build it this morning. * adiamas f***ed up... see... before i went to see the girlfriend this weekend i 'cleaned' up[ my sim dir... * adiamas doesn't really remember deleteing the files.. but they are gone so it's gonna be a few days to finish :( :-/ We have a saying: "do it properly, or do it many times" :-) oh.. i did it properly... if deletion was our aim... and now you will do it again lol yup oh.. btw.. was it cool that i went to function pointers on the sim? just seemed to make more sense yeah, its good I just got the MAS DMA working... *cool* * Zagor feels Linus breathing down his neck again... badger: what is that header "unistd.h" ? btw, sorry for *always* changing the letters b and g ... hehe We have a serious problem with the project... whats that? calpefrosch: I don't think that should be included Apparently, Zagor pretends that he has a life, seriously slowing down the development!!! :-) calpefrosch: as that is a unix-only include file LOL hehe hahaha * adiamas pokes Zagor oh.. i had an idea while driving my 2.5 hrs back home.. I knew you'd see right through me :) i was doing random through one of my playlists... random thoughts are good :-P adiamas: you have a 2.5h commute?!? and realized it would be _really_ cool if i could hit a button and have the current song added to a 'favorites' list Bagder: It's in the ID3.c. Could you remove it, or do you need it there? adiamas: great idea Zagor: nah.. only to see my girlfriend. ah i figured.. this 'favorite' list could just be a dynamic playlist... adiamas has a life too? calpefrosch: *gone* because with 1000+ plus songs it _sucks_ to have to manually pick out my favs... Just delete the other ones. :-) Bagder: tnx Linus: a girlfriend 2 hrs away, a job, ride for my ambulance core ... _And_ i still find time to hang out here and be unproductive and sarcastic... :) * adiamas smiles at linus * Linus smiles * adiamas thinks your just jealous cause i wrote a cheap little screensaver... :) * Linus tries to have a life too...wife and kids! Linus: that's a life? ;) congrats... :-) how long you been married mate? Almost 2 years 2+ here hehe im just starting to get to that point... friggin distance is killing me --- Tumm is now known as coyote- umm... any chance of having the DISPLAY value in the sim makefile be a bit more useful? someone who doesn't know the project won't know which is which.... CHARCELLS vs BITMAP well I didn't wanna use the actual target-defines how do you mean? we use ARCHOS_PLAYER and ARCHOS_RECORDER for target builds ahh.. i got ya... but they imply more than just different LCDs maybe we can just add two big comments in the Makefile for now, explaining which is which well.. its only the recorder that has BITMAP right? yes well.. instead of HAVE_LCD_CHARCELLS .. what about HAVE_RECORDER_LCD thinking about it... nods ARCHOS_PLAYER and ARCHOS_RECORDER would probably be best then we'll be forced to include "config.h" properly hmm.. well.. id rather stay the way we are then for now... anyway, those defines are documented in firmware/CONFIG * adiamas nods Bagder: do we use harness.c or app.c for the sim ? calpefrosch: remove the harness.c completely from the sim it isn't meant to be there app.c is the app main entry ok * calpefrosch have to go to a meeting. brb --- calpefrosch is now known as calpefrosch|meet could someone explain this to me? $(CC) $(CFLAGS) -c $< -o $@ i know cc and cflags -c im good -o im good but on a makefile i forget what $< and $@ are... $< is the file on the right side of the : in the rule-line above okay... and $@ is the file on the left side of the : above ahh.. okay.. thanks Actually, $< is the file that triggered the rebuild, isn't it? no, that could be a .h file So what is $< when there are two files on the right side? the first one I think the first, I think :) OK. And $* is all of them, right? ummm 'info make' ;) im just trying to find a way to clean up the sim makefile i have to believe there is an easier way to build the app obj files w/o having to do : filename.o: $(DIRPATH)/filename.c $(CC) $(CFLAGS) -c $< -o $@ 10 times this klez virus sure is annoying heheh whats the prob? it mails got ya I've received >50 mails with klez yeah it fakes From: line, and it fakes Subject: and it attaches the virus --- calpefrosch|meet is now known as calpefrosch as far as menus go... for sim and target.. should we assume that all menu items will be the same size... and if the user wants to increase the font, they do so for all menus? for now, sure adiamas: oh yes, certainly k hm, the cvs page doesn't mention the cvs-commit mailing list right is that one mentioned on any page? in emacs.. how do i open a fresh buffer? i don't think so, only in mail like the scratch buffer.. adiamas: C-x b adiams: control-x b thanks not switch.. i know that one... to create a clean unused buffer... enter a name of a non-existing buffer ahhh k *plopp* Intuitive, right? :-) hehehe sure im looking at app.c wouldn't it make more sense to have #include "tree.h" and #include screensaver inside the #ifdef HAVE_LCD_BITMAP? it depends, the tree.c functionality should probably work on the Player too the screensaver though, could very well be inside that #ifdef k... ill fix that then Zagor: how do I figure out the attributes of a dirent struct? Can I see if its a file or dir there? yes, I just have to add the ATTR* macros from fat.h into dir.h ok, then I'll need to make sure the simulator supports them yup not perfectly POSIX, but I think it's a fair trade-off I agree umm.. what do we call a menu item? is the generic term 'item' okay? more specifically 'menu_item' menu_item sounds nice :) k... hehhe okay.. its offical.. we have the first stages of a menu api goodie once i commit, im going to ask you to look at the menu.h file and see if the functions make sense w/o comments. ok the struct name shouldn't use uppercase at all? what's the difference btwn put_cursor(0) and put_cursor_menu_top? no, only lowercase scrolling menus... idea is that put_cursor puts you at the top of the menu all togeather.. put_cursor_menu_top wold be viable menu right now there is no difference ok basically.. put_cursor() puts you _anywhere_ ill clear it all up in comments as i go... ok that and ill try for better names... i just get sick when i think of "put_cursor_to_top_of_visable_menu()" :) uh, me too... I prefer fewer functions with parameter rather than many (void) functions within reason, of course so do i... downside right now is there isnt much our menuing is doing... always easier to coalese basic functions togeather... least for me yeah, I'll give you some slack for the moment :) oh sure.. make it easy... what im trying to figure out is how to handle sub menus... what i was thinking is that maybe we keep the menu_struct we have.. but as we add submenus we just create an array of the structs... adiamas: first, can you think of any submenu we should have? and jsut reference them.. but im not sure.. sure... :games right and when we get more ppl playing with the ui, itll get more robust... options menu i ment to ask.. seeing as you'all are working in firmware... i know the idea of a 'boot' loader is not a good idea... but what about the ability for the user to have a menu to choose which version of firmware to boot with next? of course we want that rather then having to copy files too and from the archos to play with firmware exactly you could have X different firmwares nods I can't recall that anybody said a boot loader was a bad idea... * adiamas assumed.... so shoot me :) Actually, we have talked a lot about that, the project name is called RoLo for Rockbox Loader sweet... ...as candy you guys have done a fantastic job getting off metal quick.. have i mentioned that? * Zagor notes it's even in the Activity chart on the web page Thx umm.. in API we list mp3info as returning FALSE if successful huh? that can't be right. correct that's wrong ;-) k.. ive got it euhhh no it is correct okay.. so it _should_ be FALSE? ummm.. that seems really counterintuitive guys? it works that way now not really because it leaves the opportunity to return different return codes for different failures but TRUE should _never_ be a failure... unless the function is invalid_mp3info() I made the win32-sim working. how can I diff so you can work with ? If you plan to use other return codes that TRUE or FALSE you should use another return type. * adiamas nods yes, but I wouldn't have to change the checks I don't follow non-zero would mean error i still don't get why FALSE means you _were_ able to fine mp3info FALSE is boolean, this is int it's 0, not FALSE Bagder: of course. That is why you should use int nag nag calpefrosch: I correct correct 'd prefer a diff -u ... And the comment above the function says it returns "void" :-) hehe That is a rather redundant comment. it *did* Bagder: ok, but I'm a windows-user. I only have checkboxes in wincvs ;-) well.. im submitting grammer corrections now.. argue once it's in Anyone can see what the function returns which should I check ? calpefrosch: ok, a unified diff, you should be able to make one of those I guess okay.. commited is there a way to prevent dependancy checking when you are doing a make clean? calpefrosch: I'll add you for cvs commit access. what's your sourceforge ID? I have no ID yet ok I will create it now, ok? run run! ;-) am i correct that to adjust the lcd size i edit lcd-x11? or lcd-win32 respectively? what size? of the window? well. when i compile the sim for recorder.. size is fine.. when i compile fore player... it doen'st make sense to keep the ui the same size.. true dunno how to fix it.. wantted to see how to do it the size for the x11 sim window is in the screenhack.c file is that default_defaults? yep is the player the same width? just only 2 lines hight? no it's 2 X 11 letters but there should be room for some icons too, I think --> linuxstb (dave@dsl-212-23-31-215.zen.co.uk) has joined #rockbox I'm not really good on the player lcd hey Dave Hello everyone. k.. then ill leave it for tomorrow.. i need sleep bad... night all Hi! welcome dave night adiamas * adiamas waves --- You are now known as adi|sleep Zagor: my ID is 536571 I need the username calpefrosch what else ;-) :) :-) now you're added. and Dave, you've been added too. Thanks. I plan to work on the directory browsing a little more today - mainly to implement scrolling. goodie nice it _is_ a bit limited right now, indeed ;-) tnx Has anyone thought about an API for actually playing audio files? linuxstb: we'll change the is_dir() stuff to instead use the dirent struct as soon as we know how Zagor will feature the atrributes Agreed about is_dir() - that was just a quick hack by me. oh sure, it works fine for now linuxstb: not yet. it'll be something simple, since the playing is handled by DMA and a "feed" thread I have a little experience of using libmad, so would like to implement mp3 playing in the simulator. hehe sure Zagor: can you help me with the setup of cvs? Linus: you should add the timer interrupt code to the main firmware dir. we need to use sleep() calpefrosch: not for wincvs, I'm afraid. follow the wincvs link on the cvs info page. ok It's in kernel.c Linus: yes, but it's not started kernel_init() ok, good You will have to lower the interrupt mask with set_irq_level(0) after kernel_init() And I use system.c for the interrupt vectors so the current system.c is good? Works for me ok Actually, it's quite handy with the weak symbols when developing blaeh :) Linus linuxstb That way I don't need my own system.c in every test directory --- Zagor is now known as Zagor|lunch --- Linus is now known as Linus|pizza wow pizza * Bagder starts to get hungry too Re MP3 api: I think we need a "play queue" in memory of the filenames of the next 3 or 4 tracks to play. The MP3 thread "just" plays files from this queue, and the UI thread modifies it. or perhaps the mp3-playing thread have direct-access to the playlist/queue we work on What is the difference? I'm just thinking what happens when you select "back" if the mp3-thread can peek in the list, it just picks the previos tune and plays it if we'd feed the thread with a special queue to play, we'd need to take care of special situations I would like a simple MP3 thread that is independent of the way songs are selected. off for lunch --- Bagder is now known as Bagder|eats I was thinking what information the MP3 thread needed, and the minimum is just the next n tracks - enough to keep the play buffer full. play buffer full? how much ram do the recorders have? more than 2MB? Bagder: Do you know why I always get this message: cvs [server aborted]: "commit" requires write access to the repository if I want to commit? The cvs server has a list of people with commit access, which I assume you are not yet part of. --- Linus|pizza is now known as Linus Zagor wanted to put me on the list Zagor tells me that you have checked out anonymously. You have to check out everything -ext but backup the entire tree first. :-) --> alkorr (alkorr@srs03v-1-73.n.club-internet.fr) has joined #rockbox ahhh, tnx Yo alan! yo Linus, not eating ? full of pizza right now lol well it's time i have all commited <-- Zagor|lunch (~bjst@labb.contactor.se) has left #rockbox see you later <-- Linus (~linus@labb.contactor.se) has left #rockbox <-- alkorr has quit (Client Quit) --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox <-- Zagor_ (~bjst@labb.contactor.se) has left #rockbox --> Linus (~linus@labb.contactor.se) has joined #rockbox --- Bagder|eats is now known as Bagder Bagder: I'm just commiting the win32-sim, I hope it is everything ok cool calpefrosch: Why both logf() and debugf() logf was missing. I don't know where it should be... where is it in X11 ? I think it should use debugf() instead My question was, why do we need two functions for the same thing? or do they? it isn't really clear to me what debugf() does in the simulator The simulator doesn't have a console? I think so but what's mem2hex and putpacket ? hmmm, they use logf e.g. in play.c and the debugf in playlist.c Strange. if I start the simulator in VC++ I get a console output... and the simulator don't set DEBUG either, not now at least Bagder: those functions are for GDB communication so what does debugf() do when used in the simulator? Whatever you, the simulator writer, wants it to. Preferably write to a console. sure, but it doesn't do that now, does it? I'm sorry, but unfortunately I have to do "real" work now. :-( see you later oh Bagder: I guess not. there are two implementatins * calpefrosch is going to work see ya calpefrosch CU! --- calpefrosch is now known as calpefrosch|work Linus: do you use the DEBUG define for anything when you build your stuff? No, not that I can remember. ok I think that debug.c could contain code for all platforms. I think so too I'll make it Or is it kludgy? debug() it is debugf() even Greato! <-- Linus (~linus@labb.contactor.se) has left #rockbox --> Linus (~linus@labb.contactor.se) has joined #rockbox --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox --- Zagor_ is now known as Zagor * Bagder got no post-lunch coffe, the machine is broken. aaaaaah! * Zagor is fetching his third after-lunch coffee :) * Bagder hears that someone has fixed the machine and ruuuuuuns turned out to be a rumour, had to walk down some stairs to get a cup I thought you (zagor and Bagder) are working in the same company... we're consultants, so we're hired to different customers In fact, I and Zagor are at the same customer right now. aha, I see. We are also working together with some consultants so we don't actually meet in person that often Zagor: thanks for the atributes, I'll try to make the simulator use them tonight or so nice I make a scrolling feature for the simulator Should I check in ? scrolling feature? scrolling thru the dir... And I'll just go and kill myself. I HATE THAT F***ING MAS DEMAND PIN! Linus: what's the problem ? you can tell me, I listen to you :-) The problem is that the Demand pin on the MAS is going up and down like mad in the beginning of a song. And the Player only generates an interrupt when the signal goes low so I have to poll it to find out when it goes high very often... too often... indeed -o where is the input-pin at the SH1 ? The Archos firmware initialization bears some evidence of them using a timer to start the DMA. I may have to disassemble it to see what it does. It's on IRQ6 gotta go, might pop by later <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox CU! I would like to commit tree.c with the changed scrolling. Can anybody test it with X11 ? Yes - I can check tree.c with X11. mailto dave@dchapman.com --> wavey (~wavey@host-54.valtech.co.uk) has joined #rockbox i'm off. see you! <-- Zagor (~bjst@labb.contactor.se) has left #rockbox linus: congrats with the mas :) Thanks But it is really hard to write a driver for it that doesn't hog the CPU... --- coyote- is now known as Tumm <-- Linus (~linus@labb.contactor.se) has left #rockbox --> edx (~edx@pD9EAB563.dip.t-dialin.net) has joined #rockbox <-- wavey has quit (Read error: 104 (Connection reset by peer)) Is there a maximum number of files/subdirs in a directory for FAT32? Is the root directory different? linuxstb: It's working perfectly... please commit the changes. Maybe it would be too slow at long dirs, because we search everytime from the beginning. We need to read the whole directory into memory for two reasons - caching and sorting OK - I'll commit the new scrolling tree.c tnx --- calpefrosch|work is now known as calpefrosch|away new tree.c committed <-- calpefrosch|away (~calpefros@62.52.174.30) has left #Rockbox --> Linus (~linus@labb.contactor.se) has joined #rockbox --> calpefrosch (calpefrosc@p5082C6D8.dip0.t-ipconnect.de) has joined #rockbox yo yo, brb --- calpefrosch is now known as calpefrosch|brb --- Linus is now known as Linus|away --- lou_soyur is now known as adi|work <-- edx has quit (Read error: 110 (Connection timed out)) <-- calpefrosch|brb has quit () <-- linuxstb has quit ("using sirc version 2.211+KSIRC/1.0") --- Linus|away is now known as Linus <-- Linus (~linus@labb.contactor.se) has left #rockbox --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox --> calpefrosch|away (~calpefros@62.52.174.30) has joined #rockbox morning bagder --- calpefrosch|away is now known as calpefrosch morning all morning --- calpefrosch is now known as calpefrosch|away --- ChanServ gives channel operator status to Bagder --- You are now known as adiamas hi adiamas --- ChanServ gives channel operator status to adiamas hey Bagder guess what? tic tac toe rocks? ;-) hehe not yet but... --- adiamas removes channel operator status from Bagder hehehehe --- adiamas gives channel operator status to Bagder just felt like being silly :-) * Bagder walks away to check if the coffee machine works today --> Linus (~linus@labb.contactor.se) has joined #rockbox morning Linus mooning :-) --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox --- Zagor_ is now known as Zagor hey all The feeding of the MAS with DMA works splendidly! you rock Linus I do! :-) But I think those Archos guys save a penny too much when designing the hardware saved any ata/fat32 mp3-reading in sight? Zagor "I Have A Life" Stenberg is lazy I suggest we replace him hehe maybe we can just yell at him a lot until he improves? * Linus screams at Zagor! Linus why do you say that about the archos guys> just look at the player's LCD i mean come on i saw wristwatches 5 years ago far cheaper with more flexible displays :p * Bagder added common/strtok.c adiamas: I say that because they haven't connected the inverted demand pin to an IRQ so I have to poll the damn thing ever millisecond every okay.. im very much not a hardware guy, so what is an 'inverted demand' pin? --> wavey (~wavey@host-54.valtech.co.uk) has joined #rockbox wavey the man! howdy :) * wavey been gamecubing all weekend ;) hehehe I hope you don't mind me messing in your code in the mean time ;-) no problem :) okay, ill ask.. bbl why did we have to rewrite strtok? * wavey waves i know there is a recursive one... wavey: I made the randomiser a lot faster, I changed the read-indices function and I made your test program build and run on my linux adiamas: because strtok() stores data in a static area Bagder i know.. but there is a strtok_r already writen in newlib? its part of the standard library in new lib.. that i don't know :) strtok_r() is not standard unfortunately that's why I have that code written in the first place ;-) thats why i asked :) :-) newlib's strtok_r uses a thread context so bagder's is better Bagder: why are you including both stddef.h and string.h in strtok? good q, remove the stddef.h it was just ripped out from my curl sources <-- wavey has quit (Read error: 104 (Connection reset by peer)) I think the thread context is a good thing. Why is it bad? it's unnecessary how does it manage "thread context" ? The context switcher (thread.c) assigns a new "current" pointer when it switches context All newlib functions with static data stores data in the struct pointed by "current" (or whatever it is called ok, thanks ...which means: a) we have to init newlib for each thread and b) we have to swap newlib context doesn't it store errno there too? Yes it does Zagor: yes so we'd need to watch out for that Indeed errno is not as much a problem, since we're unlikely to call a newlib function, yield, and then look at errno The called function itself may yield, however possibly, I was more thinking about that it stores errno relative the current and thus it needs to be somewhat correct Linus: no no newlib functions call yield() True umm... the chars shown in the char maps on the web site... how can i use those? this falls into the "neat but not immediately neccessary" category adiamas: the player chars? yeah. just display them ummmm how? lcd_puts() ? we need to convert them to our iternal format lcd_puts() is still using Gary's fonts well, not just chars even for player simul i feel like playing with sliders ah, you are talking about the simulators for volume control... ...we should add two mini icons for tracks and dirs for the dir browser on the recorder huh? browse a dir with your recorder whatch the little icons to the left of each row got ya... that's what im trying to figure out.. how to get stuff like that form the map currently, we'll have to add the hex-data manually for new widgets ;-) ack... the chartables.c file in firmware.. what does that cover again? fonts the three different fonts only fonts? yes Gary's code package had a widget.c file with some additional ones yeah.. going to look at that now btw, including the simulators, we now have 14341 lines of source code in the CVS ;-) hehe cool one more time.. what was the command to remove extra cr-lf? tr -d '\015' < infile > outfile ty does lcd_bitmap work? yes lcd_puts() uses it extensively k wasn't there a size limit? ah yes it only handles 16 pixels height, I believe holy christ.. finally got this working and the thing is huge :) --> wavey (~wavey@dlan1431.dircon.co.uk) has joined #rockbox --- Bagder is now known as Bagder|eats <-- Bagder|eats has quit (Read error: 54 (Connection reset by peer)) --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox * Bagder fixes the simulated readdir() to return a dirent struct as the common/dir.h defines Rockers But maybe it should be the other way around? ? you mean that the target readdir() should return a struct as the unixes use? Exactly first off, the dirent struct seems to be different on various systems, but it also contains lots of crap we will never need/use OK. Just a thought yes, would've made simulation easier ;-) hehe. I think the target code should be the easy one... now, shall I make the code use this new stuff and thus break the win32 simulator? B-] Of course. :-) We donn't want no stinkin' windows! since they don't even have a native readdir(), this change is possibly even easier for them Probably. gah, it broke on Solaris dirent is a pretty silly struct only one field is POSIX yes i want to avoid stat() yes I'll have the simulator use stat ;-) yes could've even make both Linux and Solaris use the same dirent code so I'll write my own instead couldn't * Bagder gets all confused in his head with this get more coffee :) good idea! * Bagder bangs his head and runs for coffee ah, much better found problem phew seems to work now on both Linux and Solaris nicers made the tree.c code use it and it worked too --> jb1081 (~jb1081@lime.ircam.fr) has joined #rockbox Hi jb! Hi Do you know if there's some hardware differences between 6000 & 5000 (got a 5000 few weeks ago) ? a 5000! seems to be only a smaller disk does that have 5GB disk`? Bagder: yep odd jb1081: where did you buy it? I mean, that they have two products that similar Zagor: Europe (France) ok. it's a bit odd that they don't mention it on their web page. i've heard people on the us get them too firmware release 5.08, not listed in archos website (last time I checked at least) but the led & scroll mod works yes I think it's the exact same hardware but a cheaper disk how much price difference was it btwn the 5000 and 6000, do you know? 6000 discontinued here, I think 5000: 180 euro 180 euro is quite a good price imho yes found in general distribution (food...) supermarket ! hehe one cheese, two packets of milk and one archos please ;-) :) oh, does the 6000 (5000 is the same, I presume) have the remote control pin in the headphone jack? yes cool. Do you plan to use it? sure, we'll use it just like today; for remote control we also use it as a serial port for debugging ok, I was thinking you use serial port by hardware mod to line in jack yes, that's because the headphone jack only has the "receive" pin sure we need both receive and transmit for debugging but it's the same serial port ok it's 9600? TTL? yes just thinking about using an old pc keyboard controller to build a remote but it need a clock, no? yeah, we've been thinking that too. adding a keyboard would be cool! :) it's asynchronous rs232 so it doesn't need a clock pin yep, but a pc keyboard is synchronous, if I remember correctly oh, ok "The frequency of the clock signal typically ranges from 20 to 30 Khz".. :( ok, not with a broken kbd.. too bad The serial port can use an external clock as well. synchronous yes but then we need to modify the hardware yes, but we have to open the box.. a broken serial mouse perhaps? Yup. Otherwise you have to build a separate kbd->rs232 converter I guess you can get one off-the-shelf actually did you see this somewhere (at a resonable price?) i can't remember seeing any in recent years, no :-) kbd->rs232? you mean a ps2->serial adapter? ok.. :( btw what headphone do you use? somebody tried a in-ear phone? PsycoXul: I understood a complete remote, but I doubt now.. yes, I use the Sony EX70. they are *GREAT* oh no, don't get him started on those again ;*) Zagor: ok, I was not sure between ex70 et Koss The Plug, it difficult to ear it before buying here it's Bagder: oups, I should'nt? Bagder: hehe PsycoXul: Yes, a PS/2 to RS232 adapter Zagor: the phone amp has no problem to drive the EX70 properly? jb1081: no, quite the opposite... Linus: i've got a couple of those laying around i think ... heh they're still out there places i'm pretty sure Do you know where to biy one? buy the ones i've got came with mice/keyboards a couple years ago i think Oh look on ebay, computer stores, google heh found one but a little too big.. :) http://www.iptel-now.de/HOWTO/PS2SER/ps2ser.html --- calpefrosch|away is now known as calpefrosch hi, just read the lines of today... hehe I just posted to the list too about it and I saw some stuff like "stinkin windows" ;-) * Bagder points at Linus and looks innocent so, I think I#ll have to fix the win32 sim ? yeps should be pretty easily done yes, but I have not much time the next days. * calpefrosch looks out in the sun * Zagor spots another have-a-lifer :) hehe we got a mail from another volountary win32 hacker just yesterday... :-) who? eh, "Christian" ;-) Zagor: you replied to him yet? not yet cool. maybe I'll look at the code tomorrow. what do you think about the "scrolling" ? Is it usefull? we need scrolling I haven't tried it yet though ;-) isn't there a scrolling bar at the archos firmware? yes there is on the left side mhhh, so we have to scan the dir first to see how much is in there I bet people will want one of those I'm afraid we'd need to do that, yes ok, we have a works meeting now. cu see ya cu! <-- calpefrosch has quit ("bin bals zurück") bin bals? ;-) * Bagder is gonna pretend he has a life now! <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox <-- Linus (~linus@labb.contactor.se) has left #rockbox --> chris1 (~flanz@62.132.155.14) has joined #rockbox hi chris! Björn by sourcfor. name is "flanz" ok --> elinenbe (~chatzilla@bgp01080511bgs.wanarb01.mi.comcast.net) has joined #rockbox hello there. chris1: you're addded the ata-sim code will only work on Images on disk, I want to add support to read realtime the Sectors from the Archos HD in the UISimulator hi elinenbe chris1: we already do that. look at ata-sim.c in test/ata <-- elinenbe has quit (Read error: 104 (Connection reset by peer)) --> elinenbe (~chatzilla@bgp01080511bgs.wanarb01.mi.comcast.net) has joined #rockbox chris1: but it's not really useful for the simulator, since it will use high-level APIs for all disk operations this line "file=fopen("disk.img","r+");" say me , the code work on a pre create disk image. yes I copy data from a real FAT32 disk into a image file and then use it for testing ok your a right. With the image your need not to connect the RockBox to the USB port while the debuging session :) how is the progress coming? the MAS DMA code is coming along, although it will not be as pretty as we hoped the ATA code is bugging, and I'm trying to figure out why and right now I'm writing open/close/read for basic file access blah pretty... shmitty... who needs pretty code? well it's not just the code, more the design. we'll need to poll a pin every millisecond. we hoped to be able to get an interrupt for that instead --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox but it's not a major problem, just a nuicance <-- Zagor (~bjst@labb.contactor.se) has left #rockbox --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox --- Zagor_ is now known as Zagor bleh, my brain hurts. read() gets too complex. drink coffee! ;-) * Bagder returns an advice ;-) hehe I'll go and make food now, see ya <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox <-- Zagor (~bjst@labb.contactor.se) has left #rockbox --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox strange happenings... --- Zagor_ is now known as Zagor ok here in Berlin/EU is al lite release party inhouse.I have to go. <-- chris1 has quit (Remote closed the connection) i'm off. see you later. <-- Zagor (~bjst@labb.contactor.se) has left #rockbox --> Zagor (~bjst@as9-5-6.k.s.bonet.se) has joined #rockbox --> calpefrosch|away (calpefrosc@p5082C815.dip0.t-ipconnect.de) has joined #rockbox hello <-- wavey has quit (Read error: 110 (Connection timed out)) --> wavey (~wavey@host-54.valtech.co.uk) has joined #rockbox --- ChanServ removes channel operator status from adiamas <-- Zagor has quit ("bye") <-- wavey has quit (Read error: 110 (Connection timed out)) --> wavey (~wavey@host-54.valtech.co.uk) has joined #rockbox <-- elinenbe has quit ("ChatZilla 0.8.6 [Mozilla rv:1.0.0+/20020507]") <-- calpefrosch|away has quit (Read error: 104 (Connection reset by peer)) <-- wavey has quit (Read error: 110 (Connection timed out)) --> Linus (~linus@labb.contactor.se) has joined #rockbox --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox evening afternoon :) :-) moo * Bagder scrolls in the directory browser... wooo ;-) --> quiquee (~quique@217.8.197.89) has joined #rockbox hey hello everyone its a silent period in here right now --- quiquee is now known as wannabeuihacker * Bagder giggles I have seen that the menu stuff assumes complete separation of player and recorder yes, but I did that mainly to get a player version that compiled and built for the simulator that is I added the menu functions and got a menu system for the player cool mostly copy and paste I guess it is the best solution for now yes we'll see if we can merge stuff later on or if we'll keep them apart but the button right of the player is not working, any clue why ? I mean , when using the simulator the x11 one? yep weird, it should work when you get your code in, I can have a look I am trying to find out , how do I flush printf output ? fflush() ok I am doing something wrong then. right button works my code doesn't heh fflush cores :? anyway I am trying to get used again to emacs and c fflush(stdout); it should be hmm , thanks. or whatever stream you wanna flushs I didnt use it for 6 years now this is a perfect way to get back in shape ;-) do you use emacs ? would you happen to know how was to run make withing emacs without a need to open a shell buffer ? yeah... M-x compile shit , I thought it couldnt be that easy meta-s is a common bind for it... M-x global-set-key ;-) ok , it insists in running make from the ./uisimulator it would be a good idea to add a Makefile there right, it runs from the dir of the buffer you start it from I will have to read again the gdb info -wannabeuihacker/#rockbox- panics * Bagder tries to paint a little icon for the files in the dir browser unfortunately, gimp doesn't output our format natively ;-) ok, it is working I plan to create a menu structure that is read in the future from an external file so people can customize it to whatever they want just an idea to start playing with it no plans to add an xml parser in the firmware, right ? ;) I don't think that'll be necessary, but we'll see hehe I am having a lot of fun with this, an excuse to go back to the roots. Thank you guys I think I speak for most of us when I say that we enjoy this a lot too it is contagious . going to have a nap now. cu soon see ya The ATA code sure is full of bugs...I'm killing them by the dozen...:-) * Bagder hands Linus the fly squasher How can you do sleep(HZ/500) when HZ is only 100???? :-O haha short sleep ;-) And it's full of return -1 for error when the code checks for 0 heheheh Well, it's starting to work now. I've read the boot sector. on target? Yup rock I wish Zagor was on the channel. hm, I think I have a little perl script now that can "fix" an xbm as saved by gimp, to become "rockbox" formatted bango! heh works now, what does a symbol for a directory look like? ;-) Go figure disk_init() finds the correct root partition now. but readdir() fails miserably... * Bagder commtted his little icons linus = sleep(until(6am)); bye night <-- Linus (~linus@labb.contactor.se) has left #rockbox <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox <-- jb1081 has quit (Read error: 113 (No route to host)) * adiamas pokes you all in the forehead ouch wuss yea im a wuss has anyone figured out why the remote does not work with the recorder? --> calpefrosch (~calpefros@62.52.174.30) has joined #Rockbox morning evening :) oh had a nice day ? ;-) heheheh quiet so far... but im on call till 0700... so ill tell you tomorrow :) it's tuesday on your side, right ? * calpefrosch goes for breakfast --- calpefrosch is now known as calpefrosch|eats yeah... its actually wends now. --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox hey ho <-- wannabeuihacker has quit (Read error: 110 (Connection timed out)) --> wannabeuihacker (~quique@217.8.197.187) has joined #rockbox <-- PsycoXul has quit (Read error: 104 (Connection reset by peer)) --> PsycoXul (psyco@adsl-63-205-40-140.dsl.lsan03.pacbell.net) has joined #rockbox hey Bagder --- calpefrosch|eats is now known as calpefrosch mornning calpe The sim is running, where has it changed ? ahh, I have to fill the attribute exactly and size too preferably --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox howdy Zagor_ --- Zagor_ is now known as Zagor i'm puzzled by linus' ata changes. they don't work for me odd he said he had a few problems but he could read a few things on target i guess he'll arrive soon yeah, he got to bed early :-P surely before 1 am... normally, I'd boo him too :) hehe I kinda like being the one who says good night when he leaves and good morning when he arrives back again ;-) hehe btw, we probably will end up needing a malloc() sooner or later, don't we? yes I thought I'd have a go at that we get a malloc with newlib, if only we supply sbrk(). but I like your malloc, so go ahead :) hehe uh wannabeuihacker has a little menu app written for the player now he said yesterday ah, cool calpefrosch: any plans on making the win32 simulator work as a Player? * Bagder runs for more coffee ok, sim is fixed. ain't my icons beautiful? ;-) I love it Bagder: yeah, I can try to made a player. But, as you know, I'm not good in VC++ It would last a bit well, maybe some of the other guys could help out on that part too yes. Another question... I get an error in id3.c that header is not initialized which line? 351 that's a good warning it should be set to 0 do you commit in cvs ? I'll do it, I fixed another thing too Bagder what is is the menu app that wannabeuihacker worked on .. do you know? it's written specificly for the player, that's about all I know the one we have now is recorder-specific nod btw.. bagder.. perhaps its time to start getting togeather an icons.h file? yes where did you put your icons? both a icons.c and an icons.h it's in the tree.c file they're im adding icons.h right now. feel free to give my icons better names should we rename "fileimage" to "fileicon" ? yes those are 6x8 right? * Bagder nods will the player use icons? it a limited extent in a okay... just wondering if i was going to #defin around it but hardly the same ones nods anyone have an issue with the function name "get_6x8_bitmap_icon" or would you rather refernce the array directly? I prefer referencing the array directly no reduce all overhead to reduce that case.. any prob is all 6x8 icons are in an array with the subscripts as enums? Dave posted a good idea haveing X # of icons is annoying... havn'et seen it yet which part of the idea did you like? storing short names in memory well.. i don't know if i agree with his "just 16 chars" and the short name... dunno if i like that either. why not? i store my mp3s as "artist - song name.mp3" it's a memory internal thing so, stuff like "Les miserable - Curtan call" or whatever.. it doesn't help me on the dir listing right, but we haven't solved that issue yet right... i guess i just don't get what his suggestion is attempting to solve I'd prefer having a way to strip off the current directory's name from the local files or something you need to check the tree.c source to see what is aiming at ;-) what he is it currently stores the full names in memory and displays the 16 first letters i agree that we strip the dir structure..t hat just makes sense... but limiting the filename that cane be displayed.. dunno if i agree with that. thus, the displaying and the storing in memory are two parts If so, the buffer in tree.c just needs to store the short name (for accessing the file) and the first 16 characters of the full name (for display on the screen). 16 chars is _not_ enough space for display thats what im getting at I get it but the full name shouldn't be needed either in all cases --> wavey (~wavey@dlan1431.dircon.co.uk) has joined #rockbox anything internal.. hell we index however we want.. then when do you get it? if not when you do a dir listing? the full name stored for diplaying doesn't need to be the full file name it could be modified to look better define "look better" stripping off redundant parts for example but "Beat on the brat.mp3" is good.. that i agree.. like I have "artist/artist - tune 1.mp3" but if you stop me from displaying my file "Les Miserables - Javier at the gates of the city.mp3", ill slap you silly :) then showing the dir in 'artisy' could cut off 'artist -' for me hehe well you still can't show that full length on each line i agree.. any directory struct stuff should not be shown sure you can.. it's called scroll :) you wanna scroll all lines? no.. only current selected line.. as you page through... right but if you only get the ful name when you cursor over it... but then you don't need to store the full name in memory you are going to be spinning up and down the drive constantly sure you do ah true now.. we can make alterations.. say.. ".mp3" can be dropped.. and all directory structure... that i fully agree.. oh.. and another suggestion.. you should not be able to try and play _non_ mp3 files i use my archos for data backup as well.. and my gf tried to 'play' "Oracle8" hehehe hehe play didn't like that much was it good? ;-P id rather display "Not an mp3" or something when you select it... and it's easy enough a check to do when you select a file to play Bagder wanna do me a favor? sure could i mail you this icons.h and tree.c for the uisim.. let me know if the idea makes sense? it works.. just dont wanna set bad precident sure, fire away so no reasont to check it in i think it should do something specific to the filetype whenever it can, even if its something simple, and some simple thing for whatever unknown things eg you "play" a text file it shows it to you and if you play a HTML file ... *g* heh * Bagder grins yes so when are we getting the modem driver for the serial port? lol uh, that'll be after I'm done with the apache port hehe bag, sent gotem * adiamas thinks you should all read quesiton # 17 on the FAQ adiamas: I like that approach, I only have some nits oh yeah btw Q18. I don't see you mentioning ogg files on your list of ideas. What about supporting those? that question kinda seems odd when right within the same page you can see - Support playing of other files types (ie: Ogg Vorbis support) in said list of ideas adiamas: the array should be in the icons.c file, and the icons.h should have the enums and the array "extern" well.. thats why i was planning on having the calling function to get the array having externs all over the place is nasty but thats just my opinion... ill set it up your way and we can argue later :) :-) and i couldn't come up with a better name then that :) having extern in .h files is not nasty only in .c files your right... but if we do it continuously... your going to get files with 20+ externs... and name collisions are bound to happen... ive had them happen already be more creative with your names well if they're extern they're part of the API and should be named apropriately i'd show you some of the ones i come up with, but you'd probably be offended :p zagor i agree.. but if two ppl 'name appropriately' it doesn't prevent collisions :) sure it does how? adiamas: right, but then we smack them and joy is restored if you're using some api you don't name your vars by its external var naming conventions heh PsycoXul sure.. but were already at 14k+ lines are you telling me that you know _every_ extern we have defined already ;) no but i don't need to if i name my things in a specific manner * adiamas uses that figure because someone else poitned it out the other night. sure you do... there is no way to prevent collisions (other then slapping :) ) when you extern many of those 14K is in the simulators so it becomes either, a. know the code or b. trial and error. adiamas: I count six (6) externs currently thats all im saying take this function, for example [advance appology to young or sensitive viewers]: shit carray(ass *bigass, ass *fatass) { shit poop; three of which are the fonts for (poop=0;(bigass[poop] == fatass[poop]);poop++); return poop; } i don't think that would have collisions with much of anything what about carray? except the code ethics police :) know one else is going to use that? ;) --> Linus (~linus@labb.contactor.se) has joined #rockbox hey Linus Zagor: i agree.. its more of the 'as we go' thing... yeah well that could be named differently thats all im saying. :p it really should * adiamas trouts PsycoXul viciously. but i leave it like that anyways * adiamas trades his trout in for a nice frozen ferret i've got another one thats named much more appropriately, however i figured in case anybody wouldn't appreciate the one i did paste, they really wouldn't appreciate this one okay.. who wants to be ferreted? grrrr... can someone please make a note somewhere making a request for bounds checking? snprintf vs sprintf we should prevent sprintf somehow, so it won't link same with strcpy i suggested defining our own empty sprintf and strcpy with no args so the compiler bitches it didnt go over well :) i'm not sure linus and I just got the target ata code working anyone sits on a malloc() testing program? ;-) sweet :) *nice* so now we have full fat access on target any chance of partitioning out our unit yet or what ;) * adiamas coughs *reiserfs* adiamas: we already support multiple partitions, actually. but only fat32 so far, of course. wow.. didn't know you'll had taken that far okay, icons.[ch] created and commited tree.c updated and io.c fixed with removal of sprint, strcpy calls Zagor what do you think of a file for global #defines? stuff like the starting font size and menuing/dir y co-ordinates i feel this could be located in the ui code rather than some global file, no? well, thats what i mean... prob is, what file would you put it in only thing i can think of would be app.c why not? start with it there fair enough deal with that tomorrow.. need sleep night al --- You are now known as adi|sleep night! g'night, even Zagor: go work on the read() !!!! i'm on it * Bagder works on malloc() Bagder is a nice person * Bagder walks to fill up his coffee cup --> Zagor_ (~bjst@labb.contactor.se) has joined #rockbox <-- Zagor has quit (Read error: 104 (Connection reset by peer)) --- Zagor_ is now known as Zagor the malloc system will benefit to get adjusted to our use cases whatever they are yup anyway, this seems to work pretty good now, and I can run all sorts of test cases against it --> alkorr (alkorr@srs01v-3-66.n.club-internet.fr) has joined #rockbox hi alan! hey alan hi Zag', Bagd' what's up ? the usual all sorts of poking in sources all over Linus and Zagor are doing progress on the ATA pokemen :) zagachu and linuchoo adding interrupt in ata ? not yet we just got everything working together ok humm, my read() doesn't signal EOF. that's not very nice :) see you later see ya alan I now return -ERRNO values from file.c. anybody see a problem with that? Not at all yes You should set errno. :-) we check for -1 from read() for errors why not < 0? and lseek too because POSIX says -1 Ah. So errno is the way to go. it is yes but errno needs a thread context :( so? can't we make our own errno instead? * calpefrosch going to have lunch now so I don't want a thread context because... --- calpefrosch is now known as calpefrosch|eats because it's not really giving us anything we need you mean like errno? we don't need thread contexts for errno only because newlib forces us yes. we don't *need* errno for any other reason than to sustain strict POSIX comliance Bagder: no. Just a thread safe errno we don't need a thread safe errno heck oh? Why not? we don't need errno much at all Maybe not well, people should be used to errno not being thread safe Really? yes Well then. Go for an unsafe errno. we'll educate the masses ;-) :-) Good. What about perror()? When DEBUG compiled. strerror[] is better, imho hmm, or maybe not <-- alkorr has quit (Read error: 110 (Connection timed out)) --- Zagor is now known as Zagor|lunch --- calpefrosch|eats is now known as calpefrosch --- Zagor|lunch is now known as Zagor <-- Linus (~linus@labb.contactor.se) has left #rockbox read() is working * Zagor cheers * Bagder horrays it's a bit messy but we'll fix that closing on reading an mp3 now huh? ;-) you bet --> edx (~edx@pD9EA9824.dip.t-dialin.net) has joined #rockbox hey edx hi :) it's been a while... hehe i am working on a little secondary computer (i try to run cygwin with sshd there...) * edx has networking problems with his new computer and will be right back. <-- edx has quit (Client Quit) I and Linus discussed versioning during lunch. We came up with the idea to write a list of "requirements" that version 1.0 should fulfill. something like: Directory browsing, mp3 playback and volume control. that'd be very good we could put features on milestones makes it easier to see where to put your efforts first yes I'm writing a status report. anything you think I've forgotten? ;) uuuuuh ;-) the scrolling dir browser! adding who wrote that? you? no Dave Chapman and Stefan Meyer --> edx (~edx@pD9EA9824.dip.t-dialin.net) has joined #rockbox --- edx is now known as edx| --- edx| is now known as edx hi... is anybody of you linix guys there.. i have a question :) * Bagder is around ok.. i set up an ssh server and when i connect i get this: You are successfully logged in to this server!!! setgid: Invalid argument what is taht setgid and where is it called? gid is group id man setgid like uid is user id aha and whats not working with this edx: it is probably called by the server or something yea it is called by the server. but why :/ why? i guess its a cygwin related thing.. ? to change effective group bah.. and why does it have an invalid argument cause i don't think windows has groups... that is probably cygwin related, yes --> chris1 (~flanz@62.132.155.14) has joined #rockbox seems like cygwin would provide some sort of compatability h hi chris1 hi hmhmmhmmm... hmmm * edx will rebuild etc/passwd stuff cygwin ?? you want to use gcc on W32 edx is one of those weirdos ;-) i never said i was to build stuff with it :P just playing iwth the ssh server :) Zagor: perhaps the "app" stuff in uisimulator should be moved to firmware/app/ or something mmm, maybe since it isn't strict simulato stuff we discussed the library approach during lunch again. i'm warming up to it. ah still, the simulator directory seems the wrong place for actual code meant for target too so maybe we should make an "applications" cvs module containing all various applications instead right so do we want "app" or "applications"? :-) * PsycoXul votes for "apps" * Bagder raises his hand for "apps" too right * Bagder wants ironi's web design applied yeah, me too soon I'll just rip it myself :) :-) apps module created. create subdirs in it for each applications I'm still undecided if we should have a "player" and a "directory" directory or if we should try to make them coexist "player" and "recorder", of course in the application level, much will differ between them I guess precisely we could probably start with moving files from the uisimulator to the apps dir the firmware can load latter the Recorder code from the disc as file.I use to 90 % the Player. chris1: yeah but what's the point? we'll create separate binaries for players and recorders. edx: are you using wincvs? yes. We have some #ifdef in the player code to make the best results( smale binarys) for Rec and Player. exactly my comments applys not the the source code move talk :) * edx managed to get the ssh server working "its possible to set registers so two outputs are connected together, thus blowing the ATAPI interface" not likely sounds unlikely, yes the player means ( target player&rec ) . The function Recorder can be a addOn module, load by the firmware from archos harddisk. dynamic loading would require more code than the recording function, probably... :-) we speak of Recorder and Player when we mean the different Archos models ah, yes gotta go <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox * calpefrosch goes home bue bye bye <-- calpefrosch (~calpefros@62.52.174.30) has left #Rockbox dynamic loading can be a folder (plugins) on the root filesystem, where the User must reload a other firmware.That have no playlust support etc. That can only record, an have all the memory for this task. yes but we must write the dynamic loading code ourselves. there is no operating system. and dynamic loading is very complex it's better to have a simple boot loader that lets you select different firmware versions att startup we have plans for such a boot loader (caller 'rolo') ok, this boot loader can have return point if no firmeware select, and can call by the allready loaded firmware. no, we can not return to the ROM firwmare. once the archos has loaded our file, we are on our own. well if a firmware that the bootloader runs returns, can it continue? the bootloader i mean that's a possibility if we reserve an area in RAM for it then a firmware can "exit" by calling it also it's not impossible to have each firmware contain a snippet of code that loads "rolo.bin" and executes it heh especially games would benefit from this the bootloader would appear like an OS that lets you run programs yes cool in a way, it would look like the plugins chris talks about. but in reality they are separate applications. I think the time will show what is the best solution for this.First task is to go to R1.0. .) that can play mp3. yes --> Linus (~linus@labb.contactor.se) has joined #rockbox hi linus Yo! I'm baking with my son Zagor: is the read() working? yes i'm struggling with lseek() now I can't wait... SEEK_SET and SEEK_END works, but SEEK_CUR bugs (when used in insidious ways, that is) Ah. But how hard can that be... the problem is, of course, the sector caching and when you seek back cross-sector etc Ah. I see --- Linus is now known as Linus|baking ahhh, works again <-- wannabeuihacker has quit (Read error: 110 (Connection timed out)) gotta go. bye <-- Zagor (~bjst@labb.contactor.se) has left #rockbox <-- chris1 has quit (Remote closed the connection) <-- wavey has quit (Read error: 110 (Connection timed out)) --> elinenbe (trilluser@bgp01080511bgs.wanarb01.mi.comcast.net) has joined #rockbox --> edx|cygwin (~edx@pD9EA9824.dip.t-dialin.net) has joined #rockbox <-- edx has quit (Read error: 104 (Connection reset by peer)) heh.. got irc to work with cygwin.. finally --> edx (~edx@pD9EA9824.dip.t-dialin.net) has joined #rockbox --- edx is now known as edx| --- edx| is now known as edx --- edx is now known as edx|away hehe go you edx --> quiquee (~quique@217.8.199.41) has joined #rockbox <-- quiquee has quit ("Leaving") --- Linus|baking is now known as Linus how is the food? :) Great --> quiquee (~quique@217.8.197.154) has joined #rockbox I guess today is a bad day to check out the firmware from cvs # __________ __ ___. # Open \______ \ ____ ____ | | _\_ |__ _______ ___ # Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ / # Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < < # Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \ # \/ \/ \/ \/ \/ # $Id$ # HPATH = /usr/local/sh1/bin CC = $(HPATH)/sh-elf-gcc LD = $(HPATH)/sh-elf-ld AR = $(HPATH)/sh-elf-ar AS = $(HPATH)/sh-elf-as OC = $(HPATH)/sh-elf-objcopy INCLUDES=-I. -DBLOCK_SIZE=512 CFLAGS = -Os -Wall -m1 -nostdlib -Wstrict-prototypes -fomit-frame-pointer -fschedule-insns $(INCLUDES) AFLAGS += -small -relax SRC := $(wildcard *.c) OBJS := $(SRC:%.c=%.o) %.o: %.s all: $(OBJS) # archos.mod # archos.asm archos.elf : $(OBJS) app.lds archos.bin : archos.elf archos.asm: archos.bin archos.mod : archos.bin dist: clean: ======= sorry aboyt that --> calpefrosch|away (~hotrana@p5082C723.dip0.t-ipconnect.de) has joined #rockbox <-- calpefrosch|away has quit (Client Quit) woo hoo! I am sorry, trying to paste and copy in an xterm in cygwin/X crt0.o: In function `main_k': crt0.o(.init.text+0x74): undefined reference to `main' when trying to compile the firmware There is no firmware. Just a bunch of source files. I thought it was already something that could compile as a whole Not yet. <-- quiquee has quit () <-- edx|cygwin has quit ("Leaving") <-- Linus (~linus@labb.contactor.se) has left #rockbox should all the apps that are in the uisim be moved to the new apps dir? --> chris1 (~flanz@Bbb2d.pppool.de) has joined #rockbox can i leave a mesg to zagor ? sure.. go ahead ill give it to him <-- chris1 has quit (Read error: 110 (Connection timed out)) <-- miah has quit (Remote closed the connection) --> Bagder (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox <-- edx|away has quit (Read error: 110 (Connection timed out)) --> wavey (~wavey@dlan1431.dircon.co.uk) has joined #rockbox hey wavey so whats the story with putting apps into the apps dir? good morning adi|sleep: just that the apps probably shouldn't be in the simulator dir since they'll be part of target builds too right.. but what i mean is, should i just move them over blindly or what? btw, it is a national holiday in Sweden today, thus the calmness ;-) hehe k.. i was wondering.. what is the holiday? it is a traditional religious thing gotcha --> edx|away (~edx@pD9EA9986.dip.t-dialin.net) has joined #rockbox --- edx|away is now known as edx|breakfast <-- Bagder (~daniel@as3-3-2.ras.s.bonet.se) has left #rockbox <-- edx|breakfast has quit (carter.openprojects.net irc.openprojects.net) <-- wavey has quit (carter.openprojects.net irc.openprojects.net) <-- adi|work has quit (carter.openprojects.net irc.openprojects.net) <-- Tumm has quit (carter.openprojects.net irc.openprojects.net) --> edx|breakfast (~edx@pD9EA9986.dip.t-dialin.net) has joined #rockbox --> wavey (~wavey@dlan1431.dircon.co.uk) has joined #rockbox --> adi|work (~adiamas@pool-151-204-140-140.ny325.east.verizon.net) has joined #rockbox --> Tumm (coyote@dreamhosted.borlange.se) has joined #rockbox --- edx|breakfast is now known as edx <-- elinenbe has quit ("ChatZilla 0.8.6 [Mozilla rv:1.0.0+/20020507]") <-- wavey (~wavey@dlan1431.dircon.co.uk) has left #rockbox --> alkorr (alkorr@srs03v-7-153.n.club-internet.fr) has joined #rockbox <-- alkorr has quit (Client Quit) --> calpefrosch|away (calpefrosc@p5082C71C.dip0.t-ipconnect.de) has joined #rockbox --> ironi (irony@pce284-lr.ing-stud1.hj.se) has joined #rockbox hello **** ENDING LOGGING AT Thu May 9 12:23:28 2002