--- Log for 01.06.105 Server: tolkien.freenode.net Channel: #rockbox --- Nick: logbot Version: Dancer V4.16 Started: 9 days and 14 hours ago 00.00.23 # and it should work with source from the daily tarball 00.00.36 # if anything of this fails, there's a bug somewhere 00.00.48 # but why the cvs source is smaller with less folders? 00.01.05 Quit thegeek (Read error: 54 (Connection reset by peer)) 00.01.57 # yeah doing the mix up produced more errors 00.02.36 # smaller with less folders? 00.02.38 # If your cvs checkout does contain less folders, you didn't checkout everything 00.03.44 # i type "cvs -z3 -d:pserver:anonymous@www.rockbox.org:/cvsroot/rockbox co rockbox" as the documentation says 00.04.02 Join Stryke` [0] (~Chairman8@cpe-24-168-110-99.si.res.rr.com) 00.04.52 # does it matters if I am in another folder when i type it? 00.06.46 # well the cvs has 5 folders where the daily build has 8 folders!! Any ideas? 00.07.09 # XavierGr: co rockbox-all 00.08.28 Quit FireEgl (Connection reset by peer) 00.11.12 # yeah thanks LinusN now it is downloading all the folders 00.11.41 # and some more even from the daily build! 00.12.35 Join FireEgl [0] (Ariel@Atlantica.US) 00.12.41 # LinusN: I tweaked my new recording loop; however, saving a file from the recording screen when a fsinfo scan is needed still takes unusally long 00.13.07 # (about 10 seconds for a write() of ~70 KB, not counting the open() and the spinup) 00.13.26 # ...even when recording is paused, i.e. without interrupt load 00.13.38 # However, I observed another thing: 00.13.51 Join ehntoo [0] (~noclue2@24-177-161-77.dhcp.mrqt.mi.charter.com) 00.14.54 # The red LED blinking for 'recording standby' is rather irregular. Maybe the recording screen doesn't yield() often enough (?) 00.15.38 Quit TiMiD[FD] ("leaving") 00.15.43 # Maybe this was introduced with triggered recording... 00.16.48 # the blinking isn't "recording standby" 00.17.13 # it blinks when the trigger level is above the threshold 00.17.15 Quit ze (Read error: 104 (Connection reset by peer)) 00.17.20 Join ze [0] (ze@ca-dstreet-cuda2-c9a-73.snbrca.adelphia.net) 00.17.20 # hmm now it is downloading this 00.17.21 # U www/internals/archos_top.jpg 00.17.21 # U www/internals/bjorn.t 00.17.28 # is that right? 00.17.30 # LinusN: I don't have any trigger activated... 00.17.34 # XavierGr: yeah 00.17.35 # XavierGr: you get the lot 00.17.38 # you asked for it 00.17.53 Quit FireEgl (Remote closed the connection) 00.17.57 # so how much data is in there? 00.18.01 # maybe rockbox-devel would have been better...? 00.18.02 # in MB? 00.18.03 # I dunno 00.19.21 # So it will download all the site!? 00.19.35 # all the cvs repo, yes 00.19.43 # LinusN: Hmm; now I 'visited' the trigger settings screen once, and it doesn't blink any more.... 00.21.13 # it stopped fehww!!!! 00.21.21 # XavierGr: A full checkout is ~27 MB 00.21.43 # good it is around that 00.21.58 # 17.2 rared 00.27.50 Quit ted2005 () 00.28.14 # hooray it built fine!! 00.28.19 # thanks all! 00.28.43 # Now i will try to fix some plugins for the iriver 00.29.04 # (not that I will succeed but I will try anyway) 00.31.53 Quit ze (Read error: 104 (Connection reset by peer)) 00.33.28 # but now I got the same error when I try to open plug-ins inside the simulator 00.33.42 # rockboxui.exe - bad image 00.34.00 # did you do "make install" 00.34.39 # would it be a good idea, to run doxygen over rockbox code to get a nice api-doku? 00.34.44 # "The application or DLL C:\cygwin\rockbox-all\build\archos\.rockbox\rocks\bounce.rock is not a valid Windows image." 00.34.51 # austriancoder: yes I think so 00.35.00 Join ze [0] (ze@ca-dstreet-cuda2-c9a-73.snbrca.adelphia.net) 00.35.02 # I once tried to, but failed for some weird reason 00.35.14 # LinusN:Just type make install instead of make? 00.35.29 # XavierGr: no, after you did 'make' 00.35.33 # no, after make 00.35.42 # :) 00.36.19 # Bagder: I will try it.. 00.39.46 # i had placed a .rockbox folder into the archos folder. I just did a make install and it gave me a rockbox.zip with a .rockbox folder where I put on the archos folder 00.39.56 # again 00.39.58 # the same 00.40.00 # error 00.40.20 # its strange though because I have an old build of the simulator and it runs fine 00.42.09 # ok i fixedit 00.42.30 # I was doing something stupid probably. 00.43.20 # Thanks and sorry to bother you. I just wanna try to help. (seems difficult though) 00.45.04 Join TCK [0] (TCK@81-86-99-123.dsl.pipex.com) 00.45.13 # LinusN: I found out why the LED is blinking irregularly. 00.46.35 # It is supposed to blink regularly but can't because the recording screen wants it to blink at ~3 Hz (16 ticks on / 16 ticks off), but peakmeter_draw_get_button() runs for 0.1 sec (10 ticks) 00.47.31 # The 16 / 10 ticks ratio doesn't do any good... 00.48.11 # ...and it seems the peakmeter is drawing too much CPU power 00.50.28 # night. 00.54.29 # night 00.55.05 Quit Stryke` (Read error: 110 (Connection timed out)) 00.55.50 # Bagder: i have done a doxygen run 00.56.28 # and it only works fine with libflac code, because it uses a doxygen-liked commenting style 00.58.29 # doxygen :/ 00.59.00 # lets say that I want to recompile aand fix a plug-in. All i have to do is find the .c file of the plug-in, make the changes and then rebuild the uisimulator? 00.59.35 # doxygen is a very nice tool.. maybe i can play a little with the options 01.00.08 # i can't remember having seen a project that uses doxygen properly 01.00.28 # everything ends up like half-baked pdf files full of function and struct declarations 01.01.31 # let's be the first! 01.03.20 # i have seen many projects using doxygen successfully 01.03.30 # ogre3d for instance 01.04.50 # i can't find any doxygen generated docs on their site 01.05.49 # http://www.ogre3d.org/docs/api/html/ 01.06.17 # riiight, i'm blind 01.07.06 # looks like the same old shit to me 01.07.22 # http://www.ogre3d.org/docs/api/html/structbsp__model__t.html 01.07.25 # nice example of what i mean 01.07.30 # it isn't even annoted 01.07.34 # it's just a reformated header 01.07.52 # it is only some infos about the API.. nothing more 01.08.22 # sure, it's just that i don't see the point in generating html/pdf docs when it nonetheless just is a reformated version of the headers themselves 01.08.28 # but it can't hurt, i guess 01.08.33 # just me being easily annoyed, heh 01.08.47 # preglow: i agree with you 01.09.37 # but it makes researching easier... 01.09.37 # i'll take that as authorative proof that i'm right 01.10.02 # well, if you think so, i'd just as well read the headers themselves, at least then i know i'm not missing out on anything 01.10.31 # i think i can find my way faster with grep than clicking around in doxygen 01.10.44 # but hey, i'm old :-) 01.11.21 # it was only an idea.. i have seen a menu entry in kdevelop for doxygen.. thats all 01.11.46 # i think doxygen might be a good thing...if used correctly 01.12.00 # sure, i agree, my original point is that it never is used correctly 01.12.12 # exactly 01.12.24 # what do you mean with correctly? 01.12.39 Join ashridah [0] (ashridah@220-253-123-50.VIC.netspace.net.au) 01.13.10 # correctly == using special comments to spice up the doxygen generated pages 01.13.31 # as libflac does it... 01.13.41 Join webguest87 [0] (~51e23707@labb.contactor.se) 01.13.42 # yup 01.14.19 # so libflac does something correctly, at least 01.14.45 # is it ok for you, that my next commits will use some doxygen style documentation as in libflac? 01.14.47 # the 5249 dma controller is so farking lame! 01.14.51 # oh? 01.15.16 Join lostlogic [0] (~lostlogic@node-4024215a.mdw.onnet.us.uu.net) 01.15.31 # it attempts to detect bad register settings, i.e bad alignment and sizes 01.15.50 # for example if size == 2 on longword transfers 01.15.57 # or even worse, 3 01.16.39 # but size == 0 also triggers a config error 01.17.18 *** Saving seen data "./dancer.seen" 01.17.27 # hahah 01.17.42 # so if the dma controller is finished (size == 0) and the audio unit wants more data, the dma controller signals config error 01.18.20 # so the dma interrupt has basically no time to react 01.18.33 # is there a faster way of compiling plug-ins instead of recompiling the whole rockbox source? 01.18.48 # XavierGr: "make" shouldn't take long if all you've changed is plugins 01.18.48 # XavierGr: should rebuild automatically 01.19.23 # make inside the plug-ins or build? 01.19.28 # XavierGr: if you avoid 'make clean', and no-one's changed a header that's used everywhere, make should only compile the changed stuff 01.19.36 # XavierGr: always inside build 01.21.15 # LinusN: oh well, not the first lame thing about this cpu 01.21.27 # annoys me to death 01.21.34 # understandable 01.21.37 # but i really should start contemplating the bed 01.21.43 # same here 01.21.50 # see you later, all 01.21.50 # gnite 01.21.58 Quit preglow ("av") 01.23.22 Quit webguest87 ("CGI:IRC") 01.23.56 # here is what I did: I changed some code to \apps\plugins\calculator.c i can compile everything up again running ../tools/configure and then make. But how can I compile only the calculator plugin? 01.25.42 # Gah, this is annoying :( 01.26.10 # LinusN: Do you have an idea how to check which thread or other code is eating up cpu time? 01.34.16 Quit _aLEX ("Leaving") 01.37.52 # amiconn: not really 01.38.18 # XavierGr: you don't need to run configure again 01.39.05 # in fact, you shouldn't 01.41.39 # just run "make" after you have changed the plugin source code 01.43.43 # thanks again it worked 01.44.49 # do you think that make install will change the .rockbox alone or do I have manually to delete the .rockbox directory and then type make install? 01.45.46 # make install updates .rockbox 01.46.03 # all you need to do is "make" and then "make install" 01.47.07 # yes but it seems that I have to manually delete ".rockbox" because if I dont I dont get the latest updates, i will try again. 01.50.53 # XavierGr: i don't have to delete it, and i haven't heard of anybody else having to do it either 01.52.42 Ctcp Ignored 1 channel CTCP requests in 0 seconds at the last flood 01.52.42 # * LinusN goes to bed 01.52.58 # cu in a few hours 01.53.01 Part LinusN 01.59.54 # http://img.engadget.com/common/images/2512853438673730.JPG?0.8177002958975231 01.59.59 # and now we are in trouble 02.00.03 # prepare for the lawsuits :P 02.00.29 # that says rocbox not rockbox 02.00.43 # plus there was a player called rockbox that showed up a couple years ago 02.00.48 # rockbox firmware came first anyway 02.00.57 Join Carman [0] (Poopie@modem-287.munger.dialup.pol.co.uk) 02.01.09 # yeah and i think this has been discussed before 02.01.15 # and beside all that firmware != player, different animal, no conflict 02.02.35 # hm 02.02.42 # I've seen it before. 02.06.36 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12") 02.11.12 Quit courtc (Read error: 104 (Connection reset by peer)) 02.20.45 Quit edx (Read error: 110 (Connection timed out)) 02.21.13 Join StrathAFK [0] (~mike@dgvlwinas01pool0-a216.wi.tds.net) 02.26.16 Join courtc [0] (~courtc@adsl-158-14-158.asm.bellsouth.net) 02.36.16 Quit Strath (Nick collision from services.) 02.37.13 Nick StrathAFK is now known as Strath (~mike@dgvlwinas01pool0-a216.wi.tds.net) 02.37.47 Quit ashridah (Read error: 113 (No route to host)) 02.41.12 Join MiLK^ [0] (Poopie@modem-287.munger.dialup.pol.co.uk) 02.41.12 Quit Carman (Read error: 104 (Connection reset by peer)) 02.41.30 Nick MiLK^ is now known as Carman (Poopie@modem-287.munger.dialup.pol.co.uk) 02.47.51 Join tvelocity [0] (~tony@ipa142.6.tellas.gr) 02.54.53 Join ashridah [0] (ashridah@220-253-122-117.VIC.netspace.net.au) 03.15.34 Join pabs_ [0] (~pabs@xor.pablotron.org) 03.17.21 *** Saving seen data "./dancer.seen" 03.17.39 Quit courtc (Read error: 104 (Connection reset by peer)) 03.18.03 Quit pabs (Read error: 145 (Connection timed out)) 03.20.30 Nick pabs_ is now known as pabs (~pabs@xor.pablotron.org) 03.26.59 Join amiconn_ [0] (~jens@p54BD7ACA.dip.t-dialin.net) 03.27.51 Quit XavierGr () 03.34.13 Join courtc [0] (~courtc@adsl-158-44-178.asm.bellsouth.net) 03.35.27 Quit amiconn (Read error: 145 (Connection timed out)) 03.35.28 Nick amiconn_ is now known as amiconn (~jens@p54BD7ACA.dip.t-dialin.net) 04.05.27 Join QT_ [0] (as@area51.users.madwifi) 04.13.37 Quit Carman (Read error: 145 (Connection timed out)) 04.15.37 Quit QT (Read error: 113 (No route to host)) 04.21.27 Quit webguest62 ("CGI:IRC (EOF)") 04.42.21 Join bipak_ [0] (~bip@p50885930.dip.t-dialin.net) 04.43.02 Part MoosCamaro 04.47.10 Quit tvelocity ("Leaving") 04.59.39 Quit bipak (Read error: 110 (Connection timed out)) 05.17.25 *** Saving seen data "./dancer.seen" 06.16.51 Join TCK- [0] (TCK@81-86-99-123.dsl.pipex.com) 06.17.34 Quit TCK- (Client Quit) 06.22.40 Quit TCK (Read error: 54 (Connection reset by peer)) 06.24.56 Join TCK [0] (TCK@81-86-99-123.dsl.pipex.com) 06.49.32 Quit TCK (Read error: 131 (Connection reset by peer)) 07.05.13 Quit Aditya ("Download Gaim: http://gaim.sourceforge.net/") 07.17.28 *** Saving seen data "./dancer.seen" 07.39.00 Join LinusN [0] (~linus@labb.contactor.se) 07.53.47 # linusn 07.53.53 # hola 08.12.52 # hmm, we need to rethink the pcm playback system 08.13.17 # the dma interrupt handler is too slow 08.13.34 # or we have to rethink the cpu_boost concept 08.21.25 # i take it you're boosting whenever's required? 08.21.31 # hence the boost-count stuff? 08.22.15 # yes 08.24.20 # the problem is that the cpu has to step down to 11MHz when changing the pll 08.24.36 # so it will run at 11MHz for up to 10ms 08.24.54 # and the dma interrupt handler is too slow for that 08.25.39 # and the dma itself eats up a considerable amount of the memory bandwidth 08.26.48 # so, we have no choice but to either a) run at max speed all the time (bye bye battery life), or figure out a way to make sure the dma interrupt never triggers while we need to change speeds. 08.27.32 # there's no way you can't schedule it to wait until immediately after a dma interrupt is finished? 08.29.52 # there are a number of ways to solve it 08.30.23 # if we decide to keep the cpu frequency constant, we don't have to run it at max speed 08.30.45 # we can set it to any speed we like 08.31.12 # we can also make sure the dma interrupt triggers when we want it to 08.31.26 # and that can even be "never" 08.31.59 # or we can make sure that cpu_boost doesn't collide with the dma interrupt 08.32.16 # or we can optimize the dma interrupt handler 08.33.18 Part courtc ("Leaving") 08.33.37 # potentially how far appart do dma interrupts trigger, for pcm playback? (i assume it's the audio output stuff that's asking for more stuff when it does) 08.33.43 # or are we talking disk access dma handling? 08.34.26 # the only dma is the audio dma 08.34.47 # the dma interrupt triggers when the dma block transfer is done 08.34.55 # and a block can be huge 08.35.06 # up to 16Mbytes 08.35.21 # in our current implementation, it's 32kbytes 08.35.40 # rougly 1/3 of a second 08.35.43 # i'm assuming the audio output stage has its own buffering, so it can wait while you setup a new range to output? 08.35.59 # that's the problem 08.36.11 # it's too small? 08.38.28 # the docs are a bit unclear, but if i understand it correctly, it triggers a dma transfer when the fifo is empty 08.38.52 # and then you only have a few microseconds to respond 08.43.36 # hi 08.44.46 # hi Slasheri 08.44.58 # i have found the reason for the cpu_boost hang 08.45.09 # or should i say reasons 08.46.02 # that's great, i was just reading the logs :) 08.46.41 # great that i found it, but not so great that it is a hard one to solve 08.46.45 Join oxygen77 [0] (~c1c28427@labb.contactor.se) 08.46.51 # now it's easier to make a fix as we know the reason 08.46.57 # sure 08.47.21 # the immediate fix is of course to run at full speed all the time 08.47.58 # the long-term fix is probably to rewrite the pcm playback code 08.48.24 # hmm, yes. Maybe we could really do the cpu_boost switch immediately after dma interrupt as said above so that should work 08.48.32 # i wish the dma controller wasn't so incredibly lame 08.51.03 # this is my idea: 08.51.41 # the current dma interrupt looks up the next buffer address/size using the callback 08.51.48 # and then reprograms the dma 08.52.11 # this may happen too late 08.52.52 Join einhirn [0] (Miranda@bsod.rz.tu-clausthal.de) 08.53.50 # if we instead provide the new address/size in another way, the dma handler can find it more quickly and reprogram the dma first, and then call the callback for the upcoming address/size 08.54.42 # i.e always have two globals up-to-date, next_dma_address and next_dma_size 08.55.06 # ah, interesting 08.55.33 # another idea is to set the size to a huge value and just let the dma run 08.55.46 # changing only the address when necessary 08.55.50 # on the fly 08.56.10 # this is not recommended by the data sheet but i think it could work 08.56.23 # Hmm, i think that both of these could be good solutions 08.57.03 Quit ashridah ("Leaving") 08.57.35 Join ashridah [0] (ashridah@220-253-122-117.VIC.netspace.net.au) 08.57.38 # the latter solution may be tricky, as the dma could run past the dma buffer if we don't change the address in time, or stop the audio request 08.58.18 # i think we should try the first solution 08.58.36 # gotta go away for a few minutes 08.59.06 # indeed, that might be better if we could it working that way 08.59.10 # i will try something :) 09.12.15 Quit rasher (Read error: 110 (Connection timed out)) 09.12.40 # Hmm, this looks promising :) 09.14.02 Join Harpy [0] (XH7VutxuZt@dsl-hkigw7wbb.dial.inet.fi) 09.17.00 # Slasheri: we could of course abandon my stuff with the add_chunk api and use only the ring buffer method 09.17.26 # Hmm.. 09.17.30 *** Saving seen data "./dancer.seen" 09.18.00 # anyway, i think this is the solution 09.18.28 # at least system is not crashing now when i go to the debug menu and change frequency 09.18.55 # what did you do? 09.19.19 # I need to do few more tests to be sure, just a moment 09.19.38 Join preglow [0] (thomj@s183a.studby.ntnu.no) 09.19.53 # took me quite a while to figure out the cause of the crash 09.20.35 # i can imagine 09.20.49 # retarded hardware does that to you 09.27.39 # by god, i am extremely tired of this computer fan hell 09.28.55 # watercooling is the solution ;) 09.31.20 # not a solution i currently employ, unfortunately 09.32.58 # :-/ 09.33.03 # there were in fact two problems 09.33.12 # 1) the dma is braindead 09.33.29 # 2) the dma interrupt was at level 7, which meant it could interrupt itself 09.33.59 # filling the stack in no time, overwriting iram code 09.35.51 # hahah, ouch 09.36.00 # well, does it fare better if you raise the priority? 09.36.59 # raise? how? level 7 is the highest 09.37.21 # lowering it made it behave better 09.37.56 # since level 7 is nmi, unmaskable 09.38.16 # bah, nevermind me 09.38.25 # i haven't done much interrupt level stuff 09.38.27 # LinusN: I think it's crashing no more :) At least i am unable to make it crash with music playing and dynamic cpu_boost enabled 09.38.40 # Slasheri: what did you do? 09.38.52 # LinusN: Do you think the whole add_chunk thing is unnecessary? 09.39.13 # it's a handy interface, but we can live without it 09.39.46 # I made the next buffer descriptors to global variables so those can be read immediately we reach the dma interrupt 09.39.58 # great 09.40.08 # gimme da patch man! 09.40.26 # just a moment, now there is a little glitch when cpu frequency is changing 09.40.30 # need to fix that =) 09.40.41 # probably some problem with that quick hack 09.40.43 # glitch? 09.41.08 # yes, some interrupt with the audio output 09.41.16 # badness 09.41.19 # :/ 09.42.58 # try puttin the descriptors in iram 09.43.07 # or at least the globals 09.43.16 # ah, i think it's buffering problem only 09.43.29 # should be fixed soon, i just made some bug 09.43.34 # oki 09.43.37 # :) 09.43.52 # make sure the globals are updated both in add_chunk and in the callback 09.46.04 # the problem was only that the pcm buffer data will get overwritten when the buffer reaches full state :) 09.47.01 # aha 09.47.22 # gimme gimme gimme da patch before midnight... 09.47.36 # of course, when it's midnight there? :) 09.48.02 # are you Bjorn or Benny? :-P 09.48.11 # :D 09.48.31 # * Bagder has his ABBA-recognition filter activated 09.49.14 Join bobTHC [0] (~foo@l06m-18-194.d1.club-internet.fr) 09.49.32 # mornin' folks ! 09.49.37 # howdy 09.50.48 # LinusN: Hmm, this would be much easier to fix if the other add_chunk code would be not needed (as ring buffering does almost the same thing) 09.52.01 # Ah, okay these are not exactly same. Ring buffer needs memcpy and add_chunk does not 09.52.09 # So it's good to have both of them 09.52.26 # avoiding memory copy is a good thing 10.05.08 Join b0bTHC [0] (~foo@l06m-18-194.d1.club-internet.fr) 10.05.30 Quit bobTHC (Nick collision from services.) 10.05.57 Nick b0bTHC is now known as bobTHC (~foo@l06m-18-194.d1.club-internet.fr) 10.24.13 Join Lynx0 [0] (HydraIRC@134.95.189.59) 10.26.35 Quit Lynx_ (Read error: 60 (Operation timed out)) 10.27.07 Nick Lynx0 is now known as Lynx_ (HydraIRC@134.95.189.59) 10.36.58 # * amiconn can catch the digital flow now :) 10.37.05 # :)) 10.37.06 # woooo 10.37.16 # crongratulations! 10.37.24 # s/cr/c 10.37.47 # don't let it run out 10.38.06 # no, i think 'crongratulations' is almost right, given that the company is korean 10.38.12 # maybe loose the l 10.38.18 # :-) 10.38.51 # clongraturations 10.39.02 # clonglatulations 10.39.06 # * Bagder giggles 10.39.18 # amiconn: w00t 10.42.48 Quit bobTHC (Read error: 110 (Connection timed out)) 10.45.02 # LinusN: i know it's low priority, but did you take a comparing picture between H1x0's and H3x0's LCD remotes ? 10.45.14 # nope 10.46.00 # oki 10.48.41 # just asking :) 10.51.06 # LinusN: Now i got it working 10.51.15 # wee! 10.51.31 # Maybe i will put the updated files to web (sorry, including the all in a patch was too hard to do :D) 10.51.42 # and after that i have to go to work :) 10.52.56 # Slasheri: good 11.11.34 # * Bagder builds rockbox with ccache 11.11.39 # faaaast ;-) 11.11.45 Join bobTHC [0] (~foo@l06m-18-194.d1.club-internet.fr) 11.14.16 # Bagder: c00l 11.14.48 # simply setting 'ccache m68k-elf-gcc' in the root Makefile works fine 11.17.35 *** Saving seen data "./dancer.seen" 11.19.25 # pretty neat app, that ccache 11.19.42 # yes, it "just works" 11.20.04 # I had forgotten I used it for my plain gcc ;-) 11.21.37 Join ted2005 [0] (ted@74pc225.sshunet.nl) 11.21.41 # wouldn't a properly written makefile do the same job? 11.34.29 Join XShocK [0] (~XShocK@pcp09492659pcs.nrockv01.md.comcast.net) 11.36.44 # sup? 11.36.48 # hi 11.36.56 Join markun [0] (~markun@bastards.student.utwente.nl) 11.37.17 # I've just been listening to weezer using rockbox! 11.37.28 # :)) 11.40.03 Join textchimp [0] (~chimp@ip67.net66.ipnetworks.net.au) 11.41.17 # i wonder, could the iriver ihp 1x0 recharge from USB like the 3x0 can, or is it a hardware limitation? 11.41.34 # textchimp: it's been done with third party mods to the cable. 11.41.59 # Loading the files to the filebuffer seems a bit slow (slower than the iriver firmware). Is that because we don't use dma? 11.42.40 # did someone check sound-playing code into rockbox without me noticing? 11.43.02 Quit ted2005 (Read error: 60 (Operation timed out)) 11.43.11 # I just downloaded the code from slasheri's site. 11.43.28 # which is where? 11.43.45 # so it's the cable rather than the player itself? 11.43.47 # nevermind, 11.43.51 # found it 11.44.00 # tell me where it is then 11.44.08 # http://ihme.org/~miipekk/rockbox/ 11.44.10 # is it working with changing songs without rebooting? 11.44.12 # textchimp: yeah, people just split out the power cable and graft a plug onto the end 11.44.59 # textchimp: misticriver.net probably has some examples floating around 11.45.17 # ashridah: the 3x0 seems to recharge from what looks like a normal USB cable, which comes with it 11.45.58 # the 3xx hardware has different wiring on the inside 11.46.07 # that's what i was wondering about 11.46.14 # as i say, people have done it with the 1xx by grafting a plug onto a standard usb cable 11.46.30 # one could probably do it by doing the mod on the inside 11.46.43 # what do i need to do to get this sound playing code to work? do i need to have the entire CVS head downloaded first? 11.46.44 # but that restricts you to charging and having it on at the same time 11.47.02 # * ashridah hands textchimp the sacrifical goat 11.47.52 Join leftright [0] (~5087ebda@labb.contactor.se) 11.49.03 # I dont think that using a hacked usb to power chord is a good idea 11.49.55 # leftright: it's been done before. of course, using your pc to do it may well be harmful 11.50.05 # but from a powered usb hub it probably would work 11.50.08 # it could exceed the max amperage that the host usb ports amperage 11.50.47 # leftright: yes. some systems have trouble with that. and of course, if the hardware's written properly, the current should get clamped. 11.50.56 # s/written/designed/ 11.51.00 # of course, that depends on the drivers too 11.51.27 # i'm yet to run into a situation where the battery life hasn't lasted long enough yet 11.52.07 # same here 11.52.27 # I suppose for folks that travel with notebooks its handy 11.52.37 # one less item to lug around 11.53.57 # wouldn't be that hard to build a mod for the power supply to split off extra chargers for things like phones/mp3 players. 11.54.28 # of course, you'd want to take care not to confuse the nokia one with the iriver one :) 11.55.15 # thats a great idea, I'm going to spice the DC chords to my notebooks power supply 11.55.36 # splice 11.56.06 # i'd be incliend to actually make yourself a small breakout box 11.56.30 # of course, that implies that you have some skill with building voltage regulating circuits :) 11.57.15 # the usb power supply is eegant though, just a pity that the H140's hardware doesn't support it 11.57.25 # elegant 11.57.45 # so sell it and get a 3xx :) 11.57.51 # nah 11.57.56 # what do i need to do to get this sound playing code to work? do i need to have the entire CVS head downloaded first? 11.58.58 # textchimp: gotta patch and copy stuff into appropriate places. 12.00.28 Quit XShocK (" HydraIRC -> http://www.hydrairc.com <- 100,000+ downloads can't be wrong") 12.01.05 # ashridah: do i need the latest cvs version or can i do it to the latest release? 12.01.56 # it might work on a daily 12.02.03 # usb current is very, very seldomly limited by the drivers 12.02.06 # patch is usually fairly forgiving. 12.02.20 # and not much has changed in the last day or two 12.02.38 # hell, i can use my usb midi keyboard without even having my computer switched on 12.02.58 # preglow: usb midi keyboard? what kind of thingie? 12.03.38 # what about it? 12.03.53 # it's a midi keyboard, just with an usb option 12.03.53 # heh 12.04.07 # so i should just grab the latest rockbox-devel from cvs? 12.04.17 # what do you reckon are the amperage parameters for notebooks ? 12.05.12 Join Naked [0] (naked@naked.iki.fi) 12.05.25 Nick Naked is now known as Hadaka (naked@naked.iki.fi) 12.05.35 # that depends on the laptop 12.05.52 # thought it was standardised 12.06.09 # how could it be standardized? 12.06.26 # Bagder: do you use ccache for the dailies? 12.06.30 # some IEEE spec for mobo's ? 12.06.30 # every laptop's going to use a different amount of current than another 12.07.12 # sorry I'm ferering to usb ports specifically 12.08.05 # oh. it's generally considered 500mA (per port, iirc) 12.08.09 # each usb port can give around 500ma 12.08.13 # but it's usually much more 12.08.18 # since there's no current limiter per port 12.09.08 # per default the current limit is much lower, and each device has to confer with the usb driver to use more power, but i've never seen that actually implemented 12.09.19 # no protection circuits ? 12.09.29 # preglow: no I don't 12.09.43 # there probably is a protection circuit, but it's for the entire usb controller 12.10.06 # i' 12.10.09 # arhjl 12.10.29 # i've managed to setup a ground loop with my pc and my printer once that shut down one of my usb ports as i was plugging it in 12.10.38 # Bagder: any plans to? 12.10.44 # heh 12.11.02 # I hadn't considered it, but yes, it would speed up recompiles considerably 12.11.12 # ah, its the specs for that usb controller that I'm after 12.11.19 # time to google 12.12.06 # Bagder: yeah, and if i don't remember incorrectly, build times are starting to get pretty high 12.12.19 # yes 12.12.36 # I'll just work out how configure should act to detect and enable ccache 12.12.53 # Does anyone remember why read() doesn't use dma? 12.13.23 # Bagder: wouldn't a simple flag suffice? 12.13.29 Join MoosCamaro [0] (MoosCamaro@m214.net81-66-158.noos.fr) 12.13.31 # probably 12.13.40 # Hi all 12.14.04 # i don't think configure scripts should second guess the programmer, ccache shouldn't be mandatory just because it's installed ;) 12.14.10 # but changing anything in the configure sequence means changing lots of other scripts that use that 12.14.38 # I think a command line option is good enough 12.15.51 # should be just fine 12.20.30 # hmmm 12.21.37 # ccache works equally well with heavily #defined source? 12.22.02 # yes 12.22.10 # then all is well 12.22.21 # what is ccache? 12.22.31 #