Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2007-07-08

00:00:10scorcheit encodes to a lot of things, so likely....but that is offtopic for this channel
00:00:25Jinx|xniJlol ok dude
00:00:31Jinx|xniJmine does
00:00:35Jinx|xniJthanks for the answers guys.
00:01:14amiconndan_a: Looks like I found the pll setup routine in the G3 OF. It's quite complex, need to fully comment the stuff in order to understand it
00:02:02dan_aExcellent! (That you've found it... not that it's complex)
00:02:44amiconnscorche: #rockbox topics became a lot shorter than they used to be. But even before that change they were always shorter than the #ipodlinux topics
00:03:13amiconndan_a: It involves udelays of up to 10 milliseconds in one code path i said, many channel on freenode use loooonch topics
00:03:26 Join jim [0] (
00:04:11Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
00:04:21Topic"Please read before speaking: | Go to #rockbox-community for offtopic/social chat" by scorche (i=HydraIRC@rockbox/administrator/scorche)
00:04:32Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
00:04:51jimis the list of compatible players on the web page a complete list? and, if not, where is the complete list?
00:04:52amiconnIt seems like starting at those disassemblies for days made me somewhat accustomed to the style apple's compiler produces
00:04:53*scorche wishes he could make the link shorter
00:04:59scorchejim: yes
00:05:01krazykitscorche, tinyurl?
00:05:14scorchekrazykit: they hide where it is taking you though
00:05:31Bagderjim: front page?
00:05:37jimok, maybe you can help... I went to the archos page
00:05:43jim(yes front page)
00:05:56Bagderit lists all supported players and no others
00:05:58jimand I couldn't find anything matching any items
00:06:18scorchenot sure what you mean by that
00:07:19jimI'm looking for a piece of hardware which is an mp3 player, which I can potentially install rockbox onto (that part is clear?)
00:07:28PaulJamthe archos devices rockbox supports arent in poroduction since a long time, so it's unlikely that you'll find them on the archos website
00:07:45Bagderjim: and what is wrong with the table on the front page of ?
00:08:08jimnot saying anything is wrong...
00:08:19Bagderbut if you're looking to buy, this is probably nicer:
00:08:23PaulJamjim: have a look here:
00:08:26jimI just can't find compatible devices so far
00:08:46Bagderfinding compatible devices (other than ipods) can be hard
00:09:15Bagderhm, sansas should be fairly available too
00:09:26Soul-SlayerOn the buyers guide, what do yellow highlights mean?
00:09:58amiconnThe X5 can still be found easily
00:10:18Bagdernot in Sweden
00:10:28bluebrotherhmm, how about a general rewrite rule for<pagename>?
00:11:01Bagderbluebrother: that's a good idea
00:11:38bluebrotherand if possible (but I guess this would be more work) make the wiki to also use those links
00:11:56Bagderthat's probably a lot more work
00:12:11Soap_or a rewrite rule for<pagename>?
00:12:21Jinx|xniJThat was a quick install
00:12:41Jinx|xniJI just bought this Ipod off of a forum, lets see what database finds for me.
00:13:03Bagderok /wiki/[page] works now
00:13:28bluebrotherBagder: nice :)
00:13:42dionoeanow time for the autorefresh build page! :)
00:14:00Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
00:14:13*bluebrother was thinking about some graphical build size overview :P
00:14:15Topic"Please read before speaking: | Go to #rockbox-community for offtopic/social chat" by scorche (i=HydraIRC@rockbox/administrator/scorche)
00:14:20jimlet me ask this... you know how some records have songs that continue into the next song but there's an index point in between? is it possible in rockbox to either have separate tracks that have no time gap between them, or to have one track covering the entire album and also have index points and bookmarks?
00:14:20scorcheBagder: thanks =)
00:14:32Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
00:14:43*XavierGr dances the "topic-change-dance" :P
00:14:51bluebrotherjim: Rockbox plays gapless
00:15:11bluebrother(unless the format doesn't support gapless, of course)
00:15:24PaulJamjim: and cuesheets are supported too
00:15:31scorcheprovided there is no gaps that are inserted into the track you possess, rockbox will not put a gap there
00:15:48*bluebrother never understood why people want cuesheets for compressed music
00:15:54jimso rockbox will play what's there and not gap between tracks
00:16:40Bagderdionoea: is the refresh time GMT ?
00:17:07jimI don't think the issue of whether music is compressed or not is relevent, because you can't really tell what the file format is of a tune when all you're doing to perceive it, is listening to it
00:17:15dionoeayes (but it was displayed in local time)
00:17:33Bagderah yes I see that now
00:17:37jimso if bookmarks for uncompressed music is found useful...
00:18:07jim...then the same reasoning can be applied to compressed
00:19:29jimbluebrother: wouldn't it depend on the content, the purpose and the person using?
00:20:10jimif it's music, you could have someone that just likes music
00:20:18bluebrotherwell, if you can do gapless I don't see a need for cuesheets.
00:20:23jim(this is the vast majority of use case)
00:20:43bluebrotherbut for some strange reasons people seem to be obsessed with that ...
00:20:58jimwhat if it's music (again) and the purpose for listening to the music is to learn the melody?
00:21:11Jinx|xniJHas anyone here tried the iMod?
00:21:20bluebrotherwhat's the iMod?
00:21:23Soap_the redwine snakeoil?
00:21:46Jinx|xniJwhatever they do, yeah
00:21:56jimbluebrother: you're looking at that as an obsession?? I look at it as a necessity for my job... I'm a pro musician
00:22:23bluebrotherwell, if you want to learn a tune, why do you need cuesheets?
00:22:34Jinx|xniJImod is something that some guy cracks open your ipod and re-routes an untouched audio signal to your bottom Ipod port.
00:22:39jimwhat I need are cues that I can -set-
00:22:51Jinx|xniJand it suppost to increase the quality of your music by tons
00:22:57jimbut I also need the cues from the original
00:22:58bluebrotherwell, that's something different than cuesheets.
00:23:02Bagderjim: that's bookmarks in our terms
00:23:26Bagderie storing (multiple) resume points
00:24:19Soap_Jinx|xniJ, kinda. He reroutes the "untouched" line-out to the headphone jack, _from_ the bottom port where it has always been.
00:24:32jimby my way of thinking, the way those are implemented is less important... therefore you have flexibility as to how you implement them... so for example you could use the implementation of bookmarks as a way to implement the cues in cuesheets
00:24:34jimor vise versa
00:24:56Bagderjim: they are both already implemented so that is not an issue here
00:25:01PaulJamBagder: but for that purpose cuesheets are propably better because you can easily skip around to the next previous "bookmark"
00:25:22 Join beam [0] (i=54bd7a8d@gateway/web/cgi-irc/
00:25:26 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
00:25:37saratogaBagder: is frequency scaling disabled on teh Sansa?
00:25:41jimanother handy feature for a musician is something like how workman can "split" tracks
00:25:46amiconnsaratoga: Yes it is
00:25:46Soap_oops, on 5th he does even less, he simply wires the lineout to the dock, where it always was. bypassind the traces with wires.
00:25:50Bagdersaratoga: I don't know
00:25:50Jinx|xniJsoap: whatever dude, same idea.
00:26:01amiconnBagder: Too slow ;P
00:26:06Jinx|xniJthat wasnt the point of my story
00:26:22*Bagder goes type practising
00:26:24Jinx|xniJSoap: Does that mean you have used it?
00:26:28jimand for a deejay, he might want to create very exactly timed loops
00:26:37saratogaamiconn: is there a reason for that?
00:26:46amiconnsaratoga: Rockbox doesn't use clock scaling on Sansa, and it doesn't even touch the pll
00:26:52Soap_Jinx|xniJ, no more than I would use a sugar-pill to treat cancer
00:27:01saratogai was under the impression the PP5022/5024 scaling bugs were worked out?
00:27:01Bagderjim: what are you saying here really?
00:27:04Jinx|xniJyeah, i dont think i buy it either.
00:27:05amiconnThe sansa bootloader sets it to 80MHz, and that's what it's running at in rockbox
00:27:24Jinx|xniJI am ordering a ALO Cotton Dock tho
00:27:35amiconnsaratoga: The pp5020/5022 clock scaling was analysed and fixed by me
00:27:38Jinx|xniJand a PAV2
00:27:55saratogaisn't the 5024 just a 5022 with an onboard DAC?
00:28:06*bluebrother looks at the circuit he soldered today
00:28:17amiconnJust today I learned that the PP5024 reports as PP5022 to software, and the PLL register setup looks familiar
00:28:25jimBagder: that when I hear statements like "I can't understand why people are obsessed with feature X", I tend to go, "take a look at the boundary of your thinking... see what's outside that boundary..."
00:28:26dionoeaBagder: you're missing an </html> tag on
00:28:28saratogaoh good timing
00:28:29scorcheJinx|xniJ: a cable is even more snake oil than the imod...
00:28:37amiconnI have a suspicion what the issue with clock scaling on Sansa is, but can't test myself
00:28:43Jinx|xniJIs that right?
00:28:53saratogaamiconn: want me to take a look at something?
00:29:01Jinx|xniJWhat do you suggest to connect my ipod to an amp?
00:29:13Bagderjim: you only reacted on a comment about why the use was needed, that use is still _supported_ by rockbox
00:29:30jimJinx|xniJ: that's a hardware issue... read your ipod manual :)
00:29:37scorcheJinx|xniJ: just talk about it in #rockbox-communuty for now
00:30:05jimBagder: good to hear/know
00:30:07amiconnsaratoga: I am 98% sure the problem lies in the timing setup by the sansa bootloader. The timing is way too slow for lower clock settings, and the old rockbox cpu scaling code didn't touch this timing registers
00:30:32amiconnThe new code does, and I think the svn PP5022 code will work like a charm on sansa
00:30:41jimBagder: are the devs thinking about a looping feature?
00:30:51Bagderjim: we already have one
00:30:53bluebrotherjim: we have a-b repeat
00:31:06saratogaamiconn: is it trivial to enable the new code or is some more work required?
00:31:07jimalso... can I select a specific time into a track from the front panel?
00:31:16amiconnThe timing can be extended a bit for sansa
00:31:26amiconnBut that's not critical
00:31:29saratogai can certainly try things that don't involve an intimate knowledge of how rockbox works
00:31:48bluebrotherthat's a special use case that like 0.01% of the people would use. If even that much
00:31:52amiconnlemme check
00:31:57jimbluebrother: that could be handy... and you have half of it already (in the form of bookmarks/cues)
00:31:58bluebrotherbut feel free to post a patch :)
00:32:17bluebrotherwell, I never missed that functionality.
00:32:45jimok, about the a/b repeat... is it possible to fine-tune the start and end of the loop?
00:33:09bluebrotheryou can set a and b points. That's all
00:33:26jimand once this is done to store, activate and deactivate multiple loopss?
00:33:46bluebrotheryou should be aware that whilst Rockbox is about playing music it's not about professional mixing and similar.
00:34:04*bluebrother throws in a link to the manual:
00:34:13jimagain, you're bounding your thinking... my question is why :)
00:34:38*amiconn builds a sansa build with cpu scaling enabled (just the binary) and will check the .map and the appropriate .o file
00:34:46*bluebrother doesn't get that line
00:34:48jimis the space to store rockbox on the player so small that innovations meant for pros cannot be accomodated?
00:34:50Bagderjim: we don't bound our thinking
00:34:59jimgood :)
00:35:00Bagderwe just focus at music playback
00:35:21Bagderbut everyone is free to work on what they feel like
00:35:27bluebrotherspace is always an issue. If the binary size increases, the playback buffer gets smaller
00:35:37amiconnsaratoga: I think it's dead simple to enable, but be aware that cpu scaling will make the sansa run a bit slower
00:35:47bluebrotherfor usual music playback this means more spinups of the disk, which means shorter battery life.
00:35:54amiconnMax clock in rockbox is 78MHz atm, but currently the sansa runs at 80MHz
00:36:10amiconnI am planning to go 80 soon in rockbox though
00:36:23bluebrotherand the archos devices are rather limited in terms of memory.
00:36:33jhMikeSamiconn: I saw (much) earlier that you found something out about the pll. I was way to busy to get into discussion though.
00:36:58saratogaamiconn: I don't think a 3% clock speed hit is a big deal
00:37:07jimI just saw a 30g archos (now having said that, I'm not necessarily married to the idea of an archos specifically, it's just the first manuf I looked at)
00:37:08amiconnjhMikeS: Yes, I learned something new about the PP5022 pll (we're still not handling it 100% correctly)
00:37:08bluebrotherso blowing up the binary size for some obscure features isn't really an option. Plus, you need someone who implements those rarely needed features.
00:37:09saratogahow do I enable scaling on my builds
00:37:36amiconnAnd I know the PP5024 pll must be identical, and am fairly sure I know what the problems with scaling on the sansa have been
00:37:50amiconnsaratoga: w8
00:37:51jhMikeSwhy does the sansa need to drop in speed?
00:37:55 Quit ender` (" But there, everything has its drawbacks, as the man said when his mother-in-law died, and they came down upon him for the fu)
00:37:57bluebrotherbeware that Rockbox only runs on older archos models −− check the frontpage for supported players (if you haven't done already ;-)
00:38:07saratogaamiconn: sure, didn't realize you were interested on working on it further
00:38:27amiconn[00:34:38] * amiconn builds a sansa build with cpu scaling enabled (just the binary) and will check the .map and the appropriate .o file
00:38:35amiconnSorry, cygwin takes its time
00:38:35saratogajhMikeS: the current scaling code maxes out at 78 MHz
00:38:46jimlet me ask this... are there hardware that is meant/built for rockbox?
00:38:56jhMikeSlet's max at the max :)
00:39:04saratogajim: hell no
00:39:20Bagderdionoea: any chance that javascript can be made to not use onload in the body tag?
00:40:03jimother than the apple players, is there a difficulty in gaining GPL-style access to programming details of newer players?
00:40:04dionoeaBagder: well you need to launch the update stuff... what would you like to use instead ? just inline javascript ?
00:40:19Bagderthe script that builds the build table doesn't "control" the body tag, only a subsection of the page
00:40:49amiconnsaratoga: It's indeed just necessary to uncomment the #define in config-e200.h
00:40:50dionoeawell you could call the javascript function in a <scipt > </script> thing
00:41:07jimopenacs and ruby on rails has a way to insert stuff into the head tag
00:41:09amiconnjhMikeS: We will, just be a little patient :)
00:41:30Bagderjim: that's really not the problem I have
00:41:37dionoeaBagder: like <script type="text/javascript">countdown_refresh(2007,5,30,21,48,00);</script>
00:42:00dionoeaI hope that waits for the page to be loaded (as well as the countdown.js script too) before executing
00:42:26 Join chrisjs169|afk [0] (
00:42:40jimwhen I upload a track to the player, can I also upload a bookmark list?
00:43:40bluebrotherbookmarks are just text files. You can move them (or create them) any way you like.
00:44:00bluebrotheryou need to use the correct syntax when writing a bookmark yourself of course
00:44:22jimohh, could you show me an example of the text for a single bookmark?
00:44:48bluebrotherjust create one on the player and open it with your favourite text editor
00:44:50jimthis would possbily solve the loop tuning problem
00:45:24jimI don't have a player running rockbox yet :P
00:45:33jimor even one that's compatible
00:45:41bluebrotheryou can use the simulator
00:45:55Bagderdionoea: reload now ;-)
00:46:08bluebrotherthere are pre-build binaries available at rashers website.">
00:46:18jimurl for simulator (and/or docs/intro/description)?
00:46:32bluebrotherand check the wiki page UiSimulator
00:46:39dionoeaBagder: cool :)
00:46:43jimahh, got it
00:46:52Bagderthe math seems a bit off
00:46:55dionoealets wait 44000 secs :)
00:47:29Bagderthe invoking looks fine to me
00:47:31bluebrotherbuild expected to complete in 44643 mins? ;)
00:47:49 Quit My_Sic (Read error: 104 (Connection reset by peer))
00:48:02jimare there simulators for linux?
00:48:16bluebrotheryes, but you need to build it yourself
00:48:49bluebrotherit's mostly a configure && make && make install thing
00:49:01 Join My_Sic [0] (
00:49:06jimI don't think that should be a problem (I figured out how to build gcc in about 1987 or so...)
00:49:12 Quit DC1 ("$4E75")
00:49:55dionoeaBagder: you need to use month - 1
00:49:59jimok, let's see, where's the source?
00:50:06dionoealike in the normal unix time structs
00:50:15dionoea(never understood why they did it this way)
00:50:32dionoeaor maybe you could change that in the .js to prevent errors
00:50:33Bagderright, the month part is always messing up...
00:50:48dionoealike target.setUTCMonth( month - 1 );
00:51:04Bagdernah, I modified the invoke to use the "raw" value
00:51:59saratogaamiconn: thanks, scaling seems to be working, though i'll let it play for a while tonight to see if it crashes
00:52:16amiconnNice :)
00:52:25saratogaas I recall, the sansa was already getting 12+ hours, so this should be neat to see how much it gets now
00:52:29 Join ment [0] (
00:52:36Bagderdionoea: hey, seems to work now...
00:52:41dionoeaawesome :)
00:52:59Bagderlooks cool too
00:53:38amiconnsaratoga: Let it play some music where the cpu needs to boost/unboost often (e.g. wma ;) )
00:54:00jhMikeSwhen both cores are used and the COP used to decode, the average cpu frequency can drop of course
00:54:08amiconnYou could cause some additional stress by using various plugins in parallel
00:54:17amiconnjhMikeS: yup
00:54:43amiconnFirst, it has to work at all though
00:54:48jimso there are no stable/testing tags for rockbox?
00:55:15saratogaamiconn: doing that now
00:55:23saratogaseems to be about 70% boosted for WMA
00:55:32saratogaat least my 192k stuff
00:55:36Bagderjim: correct
00:55:46Bagderwe're all testing non-stop
00:55:48amiconnYes. Just a little bit more than what mp3 causes
00:56:02jhMikeSone thing I'm not sure about is should pcm playback have it's core set explicitly or just run automatically on the core that calls pcm_play_data?
00:56:15jimdo you have an automated test suite?
00:56:22Bagderjim: nope
00:56:32amiconnThat's what I don't like about the wma codec yet: on arm its nearly on par with libmad, but on coldfire it's like 3 times slower
00:56:36zylcheBagder for president!
00:56:41 Quit chrisjs169|afk (Nick collision from services.)
00:56:54 Join chrisjs169 [0] (
00:56:59saratogaamiconn: my assumption is that its because we don't use much of the IRAM yet
00:57:24amiconnI tried putting various stuff in iram. ALl I was able to achieve was around 10% speedup
00:57:43saratogai was going to look into what get accessed and how often tonight
00:57:46amiconnMaybe I didn't put the right stuff in iram, as I really don't know what stuff is used most
00:57:56Bagderman that reload worked nicely
00:58:13amiconnBut I think that it will require to make real use of the emac on coldfire
00:58:18jimis it possible to restore the original operating system back onto a player?
00:58:20Bagder3:35 for the last build round has to be a new record even
00:58:33saratogawhat does the EMAC actually do on coldfire?
00:58:36Bagderjim: yes, the process is almost always reversable
00:58:40saratogajust really fast MACs or what?
00:58:45Bagderdepending on target and what you did
00:59:07jimso that would involve backing said original os up?
00:59:13 Quit My_Sic (Read error: 104 (Connection reset by peer))
00:59:27Bagderjim: it depends on each target, but it is usually documented in the specific manuals
00:59:31bluebrotherjim: most players have a dual-boot feature
00:59:48dionoeahum, my clock is off by 2 minutes :/ need to fix that
00:59:52bluebrotherthe manual holds all important details. The FAQ might also be a good start for a comparison
00:59:53jimso most of the time you're not actually -replacing- the original
01:00:25mentwhen replacing ipod's microdrive with CF, it's possible to write directly to CF with card reader before replacing (and thus skiping itunes firmware loading)?
01:00:31bluebrotherwell, the original is replaced in terms of what software is running
01:00:32 Join My_Sic [0] (
01:00:37ments/write/write rockbox/
01:00:50bluebrotherment: yes.
01:01:09PaulJamjim: well, often you don't want to use the OF anymore. so in a way rockbox replaces it.
01:01:38mentbluebrother: ok, thanks
01:01:49bluebrothere.g. on h100 you can completely remove the old firmware.
01:02:06bluebrotherthe usual attempt is to have dual-boot to allow the user to switch back.
01:02:20amiconnsaratoga: EMAC stands for enhanced MAC. 'enhanced' is related to the first coldfire MAC
01:02:31bluebrotherit's similar to installing linux on a windows box: you not necessarily need to remove windows
01:02:45menti know, but i was thinking about the disk mode
01:03:06amiconnIt's a multiply-accumulate unit with single cycle mul-add (and mul-subtract)
01:03:41amiconnIt does 32bit * 32bit -> 40bit + 48bit => 48bit
01:03:50amiconn4 accumulators, 3 cycles latency
01:03:51jimPaulJam: well that depends... I don't see much in the way of QA here (no regression testing is performed automatically, and no tagging of even last-known-good versions), so I pretty much expect things to go wrong at least initially... given that lack, I would want at least the option of going back
01:03:53bluebrotherment: huh?
01:04:36bluebrotherjim: that heavily depends on the original firmware :P
01:04:53jimno question about that :)
01:04:54Bagderjim: don't confuse bootloaders with rockbox itself. you can (in all practical cases) isntall a completely borked rockbox and just install another one without problems
01:05:07mentbluebrother: nm
01:05:27jimare "all practical cases" documented? :)
01:05:28bluebrotherI wouldn't want to go back to the iriver firmware ... and even less to the Apple one
01:05:41PaulJameven when rockbox used to crash often on h300 and batterylife sucked i still prefered rockbox.
01:05:42bluebrotherjim: :P
01:05:49Bagderjim: well, if you follow the manual and don't try going out on limbs on your own
01:06:06***Saving seen data "./dancer.seen"
01:06:14jimI take your point
01:06:21Bagdertrying to say "don't do xxxx" tend to make at least some people miss the "don't" part
01:06:36Bagderand 95% of those own sansas ;-/
01:07:09jimBagder: it's suggested by mind/NLP/hypnosis experts that it's not possible for the mind to process negatives
01:07:10Bagder(which can be pretty seriously messed up by users, not really related to rockbox)
01:07:25bluebrotherthen we should stop supporting the sansa ;-)
01:07:52Bagderwe should first ask the users if they can process negatives ;-)
01:08:03krazykitjim, i can't process negatives ;-)
01:08:10bluebrotheryeah, but ask it in a negative way.
01:08:12krazykitwhich is a negative in itself :P
01:08:20jimso if you say to pepple: "Don't order my product now"...
01:08:37jimguess how many orders start pouriing in :)
01:09:02bluebrothermaybe we should use double negatives?
01:09:06*scorche puts a sticky note on the ominous, large, red button that says "DON'T PRESS"
01:09:07krazykitso we should be telling people not to use rockbox! brilliant!
01:09:26 Quit Thundercloud (Remote closed the connection)
01:09:29jimkrazykit: you shouldn't tell people not to use rockbox :)
01:09:30*bluebrother starts to wonder how the hitchikers "don't panic" ever worked
01:09:51 Join Thundercloud [0] (
01:10:26jimof course we're talking here about the conscious mind... it's the subconscious mind that can't process negatives (according to the mind folks)
01:11:10*Bagder heads off to bed
01:11:30Bagderback in a few days (still on vacation)
01:11:35 Quit Bagder ("*plopp*")
01:11:39jimhave lots of fun!
01:13:10jim"You really shouldn't fail to press that big, shiny, attractive red button that's totally within your reach"
01:14:57 Join inversions [0] (
01:15:58Jinx|xniJWhat program do you use to put new music on an ipod with Rockbox on it?
01:16:16scorchethe file browser
01:17:17jimso it automounts? or becomes available as a usb storage volume?
01:17:33scorcheprovided itunes doesnt intercept it, yes
01:17:51jimwhy would it do tht?
01:17:56Jinx|xniJWhere do you put the music? All the music that ipod puts on it are in F folders with wierd titles
01:18:25scorchei put mine in a folder titled "Music", but you can put it pretty much anywhere you want
01:18:30 Quit bluebrother ("no more questions ... for today.")
01:19:04Jinx|xniJIncluding the root?
01:19:22scorcheyou typically dont want to put it in .rockbox though
01:20:18 Quit moos (Read error: 110 (Connection timed out))
01:20:40 Join DerPapst [0] (
01:20:48DerPapstgood evening :)
01:21:08Soap_iTunes can recognize an iPod with Rockbox as an iPod, and disable UMS mode if you have it set up that way, jim
01:21:35dionoeahello DerPapst
01:21:48DerPapsthi dionoea :)
01:21:52*dionoea didn't know that germany was a few time zones behind
01:22:09DerPapstit's 1:30am here ;)
01:23:16*Soap_ proposes rockbox transitions to SIT and forget about timezones
01:23:31jimSoap_: so it's a user-settable option switch?
01:23:45Soap_jim in itunes, yes
01:24:03jimUMS mode, what's that?
01:25:37Soap_Universal Mass Storage, when your iPod shows up as just another drive letter
01:27:52jimwill rockbox handle an itunes that's got UMS off?
01:28:18scorcheit isnt about rockbox at this is about your computer and it showing up in the browser
01:28:18*amiconn needs a longer taskbar :|
01:28:34krazykitamiconn, or another monitor :D
01:28:40scorcheamiconn: 2 spaces high?
01:28:48DerPapstor 3 even :P
01:28:53amiconnkrazykit: Difficult in a laptop...
01:29:10*amiconn has 16 app windows open atm
01:29:12krazykitamiconn, naw, it has a vga-out port for something!
01:29:24*krazykit has no taskbar at all :o
01:29:24DerPapstha.. nice
01:29:27amiconnYes, for less resolution than the built-in panel
01:29:48DerPapstno vga or div out?
01:29:53scorcheamiconn: a taskbar 2 spaces high with autohide on works wonder for me
01:30:04amiconnBah, I can't stand autohide
01:30:08DerPapsteven my 25MHz laptop has vga out ;)
01:30:20amiconn...for the same reason I can't stand grouping.
01:30:45amiconnDerPapst: It does have vga out, but the internal panel is 1400x1050, while my TFT is only 1280x1024
01:30:48scorcheand i assume 2 spaces high is too tall for you?
01:31:08DerPapstamiconn: better than nothing ;)
01:31:58PaulJamamiconn: you could use a programm like "trayit!" that lets you minimize windows to the systemtray
01:31:59jimso if I can put things into folders/dirs on the player device, is it possible to hide/disable them so that the items in the folders can't be played without re-enabling the folder they're in?
01:32:05amiconnscorche: I use a 2 spaces taskbar at work, where I only have 1280x1024
01:32:55scorchejim: well, you can hide them if you have your viewmode set right
01:32:55amiconnUsually 1-high is enough on the laptop, but atm I'm trying to do too many things at once :\
01:33:40 Quit spiorf (Remote closed the connection)
01:34:14jimthe basic idea, is I hand the player to someone, and by its normal operation doesn't play or show the disabled folders or contents thereof
01:34:54 Quit Arathis ("Bye, bye")
01:35:08jimso if I have say a private conversation I haven't transcribed yet, I can put it into a folder
01:35:13Soap_trusting that people don't have "show hidden files and folders" checked is silly.
01:35:40 Join ^BeN^ [0] (n=Paprica@
01:35:45*amiconn always sets that
01:36:14jimwhat is implied by a folder set hidden?
01:36:25krazykitjim, it means most windows users won't see it.
01:36:37krazykitand that all osx and *nix users won't not see it
01:36:47jimdoes it mean the player won't play the content?
01:37:10krazykiti don't think rockbox honors the hidden attribute.
01:37:28jimnot in any way?
01:37:31 Join MySic [0] (
01:37:35amiconnIt does for display, as long as you don't set "view files" to "all"
01:37:36scorcheit does
01:37:40jimdoesn't hide and does play?
01:37:50krazykitoh, it does? that seems silly.
01:38:20krazykiti guess players have lots of files that don't need to be seen.
01:38:40krazykitjim, it's still not in any manner remotely "securely" hidden
01:39:03scorchethere is no way to "securely" hide a folder...
01:39:03jimI understand that
01:39:04amiconnIn fact rockbox hides both files with the hidden attribute and .dotfiles
01:39:21saratogaamiconn: assuming I don't get any crashes from CPU scaling tonight, you think it could be comitted?
01:40:02amiconnsaratoga: Definitely. We might want to change the refresh timing for sansa though; from the bootloader it seems it allows twice as long timings as the ipods
01:40:10amiconnIt's not important though
01:44:51 Quit Jinx|xniJ (Read error: 110 (Connection timed out))
01:53:48 Part dan_a
01:54:57 Quit My_Sic (Read error: 110 (Connection timed out))
01:58:15amiconnHmm. Both SDRAMs have 64ms refresh period. Samsung states 8K cycles, but Hynix doesn't give that info
01:58:45amiconnSo either the Hynix ram is 4K cycle, or apple is just playing safe on the ipods
01:59:26Nico_PjhMikeS: here?
02:01:02*amiconn would like to know the H10 bootloader values for clcok setup
02:02:51DerPapstgood night everyone :)
02:02:55 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:09:09amiconnI can't test it of course
02:09:37Nico_PI'll try it now
02:10:31 Quit MySic ("Miranda IM! Smaller, Faster, Easier.")
02:12:20 Join karim [0] (
02:15:37 Quit entheh ("^~")
02:17:52jhMikeSNico_P: here
02:18:47Nico_PjhMikeS: you'll probably be able to explain queue_send to me :)
02:18:55 Quit nl2 ("Leaving.")
02:19:11jhMikeSmaybe :)
02:21:12Nico_Pwhen should I use it rather than queue_post ?
02:22:02jhMikeSwhen you want 1) a sender to be able to receive a value in reply to a message 2) you want to control when the message sender is allowed to run again
02:24:02*amiconn ->sleep
02:24:11jhMikeSnighty night
02:24:33Nico_PjhMikeS: does it block the sender ?
02:25:03Nico_Pis there a way to get a return value without blocking ?
02:25:14jhMikeSif there's an overlow for some reason, the sender will be unblocked and returned 0
02:25:22Soap_the "working with patches" wiki page covers the diff commands for creating a patch based upon changes to one file. Am I totally stupid when reading the diff −−help, or is there not an easy way to make one patch with all changes to all files in all directories?
02:25:39Nico_Phmm so 0 shouldn't be used as success value :)
02:25:50saratogacrap was I ever wrong in my estimates of which values should be in IRAM :)
02:26:40jhMikeSthere's no stable way to not block and get a return value
02:26:44Nico_Pamiconn: JdGorson's svn isn't responding and my code isn't in a state compatible with testing so i'll test tomorrow
02:26:57saratogasome table I never noticed accounts for ~ 50% of all DRAM accesses !
02:27:13saratogathe stuff I put in IRAM hardly ever gets touched
02:27:25jhMikeSthe block lets the thread that will receive the message run afterall
02:29:50PaulJamSoap_: doesn't just "svn diff > output.diff" in the rockbox directory do what you want?
02:30:16 Quit inversions (Read error: 113 (No route to host))
02:30:38Nico_PjhMikeS: ok... thing is I need to return a value for a non blocking call... I'll have to find a way to compute the value
02:31:01Soap_PaulJam, I'll try in a minute, thank you.
02:32:01Soap_svn diff −−help
02:32:11Soap_wrong window
02:33:01Nico_PjhMikeS: maybe it'd be worth adding info about queue_send and queue_rply to the wiki
02:33:35Nico_Palso is queue_send only available on SWCODEC ?
02:35:44jhMikeSit's currently only on SWCODEC but could be used on all
02:37:03 Quit lazka ("I'm off now")
02:37:23Soap_works, thank you PaulJam
02:38:23 Join Arathis [0] (
02:38:24Soul-SlayerAm I weird? My ears really can't distinguish a change in quality with different bitrates what-so-ever..
02:40:39scorcheeveryone is different, although your equipment come into factor there
02:40:59scorchei promise you will notice a difference between 64k mp3 and 192k mp3
02:41:00jhMikeSNico_P: for non-blocking I'd think the value would be based on whatever state things happen to be at the time. if you must wait for an operation to be complete on another thread like saying audio has completed the state transistion to "playing" then queue_send is the thing to use since yields during wait won't let the sender rush ahead. it's the thread-safe way to perform "while (!some_state) yield();"
02:41:14saratogaanother build I have to redo because I forgot to enable test codec
02:41:23saratogathis is slowly driving me nuts
02:42:24Soul-SlayerWould it save battery if I reencoded all my MP3's into 128kbps?
02:42:55 Join Rincewind [0] (
02:43:50 Join Arathis_ [0] (
02:43:56saratogaSoul-Slayer: depends on the player, but maybe
02:44:58Nico_PjhMikeS: the value I want to return from the non blocking call is pretty basic (just an ID that has to be unique)
02:45:45Nico_Pwhen I say non blocking, it's an API call I want to make post a message to the buffering thread and return an ID immediately
02:46:47Soul-SlayerI have a gigabeat I got on ebay, the battery usage is really erratic so I'm not sure if my music is all encoded in a bad way or the battery is messed up, either is possible
02:49:16jhMikeSNico_P: why can't that block? queue_reply can be called with an id in the retval immediately and the sender will run at next task switch.
02:49:50 Quit Thundercloud (Remote closed the connection)
02:50:53 Quit BHSPitLappy (Connection reset by peer)
02:50:54 Join BHSPitLappy_ [0] (
02:51:39Nico_Phmm that makes sense
02:52:52 Quit Rincewind ("Cya")
02:54:22Nico_PjhMikeS: I was thinking of it as "add a handle and return its ID"... thing is that can't block because we don't know when we'll actually be able to add the handle in memory
02:55:02Nico_PIf I see it as "add a handle ID to the list of handle IDs to add", it works better
02:56:45 Quit Arathis (Read error: 110 (Connection timed out))
02:58:40 Quit Arathis_ ("Bye, bye")
03:01:00 Quit PaulJam (".")
11:01:07***Started Dancer V4.16
11:01:07***Connected to on port 6667
11:01:07***Logfile for #rockbox started
11:01:10CtcpVersion from freenode-connect!freenode@freenode/bot/connect
11:01:11***Server message 501: 'logbot_ :Unknown MODE flag'
11:01:11Mode"logbot_ :+i" by logbot_
11:01:12***Server message 477: 'logbot_ #rockbox :[freenode-info] if you need to send private messages, please register:'
11:01:12 Join logbot_ [0] (i=bjst@gateway/web/cgi-irc/
11:01:12 Join spiorf [0] (
11:01:12 Join Rob222241 [0] (
11:01:12 Join dan_a [0] (n=dan_a@
11:01:12 Join petur [0] (n=petur@rockbox/developer/petur)
11:01:12 Join seclein2 [0] (n=scl@
11:01:12 Join entheh [0] (
11:01:12 Join lazka [0] (
11:01:12 Join Guile`` [0] (
11:01:12 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
11:01:12 Join desowin [0] (n=desowin@unaffiliated/desowin)
11:01:12 Join davina [0] (
11:01:12 Join bluebrother [0] (i=NEvzDZgH@rockbox/staff/bluebrother)
11:01:12 Join The-Compiler [0] (
11:01:12 Join bdgraue [0] (
11:01:12 Join inversions [0] (
11:01:12 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
11:01:12 Join hannesd_ [0] (
11:01:12 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
11:01:12 Join Redbreva [0] (
11:01:12 Join Dcider [0] (n=dce@unaffiliated/bspec)
11:01:12 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
11:01:12 Join Rob2222 [0] (
11:01:12 Join Bawitdaba [0] (
11:01:12 Join alienbiker99 [0] (
11:01:12 Join goffa [0] (n=goffa@
11:01:12 Join RoC_MasterMind [0] (
11:01:12 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
11:01:12 Join TiMiD[FD] [0] (n=TiMiD[
11:01:12 Join miepchen^schlaf [0] (
11:01:12 Join Llorean [0] (n=llorean@rockbox/administrator/Llorean)
11:01:12 Join Soap__ [0] (
11:01:13 Join merbanan [0] (n=banan@
11:01:13 Join midkay [0] (n=midkay@rockbox/developer/midkay)
11:01:13 Join Echelon [0] (i=ryan@
11:01:13 Join ^BeN^ [0] (n=Paprica@
11:01:13 Join ment [0] (
11:01:13 Join ansivirus [0] (
11:01:13 Join kubiix [0] (
11:01:13 Join dj`bergamote [0] (
11:01:13 Join jhulst [0] (
11:01:13 Join Jon-Kha [0] (
11:01:13 Join Siku [0] (
11:01:13 Join maffe [0] (
11:01:13 Join pabs [0] (
11:01:13 Join darkless [0] (
11:01:13 Join XavierGr [0] (
11:01:13 Join atsea- [0] (i=ariel@gateway/tor/x-c26824c4e1bd83d6)
11:01:13 Join _pill [0] (
11:01:13 Join tris [0] (
11:01:13 Join lymeca [0] (n=lymeca@unaffiliated/lymeca)
11:01:13 Join perplexity [0] (
11:01:13 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
11:01:13 Join Rick [0] (i=rick@unaffiliated/rick)
11:01:13 Join sargun [0] (n=Sargun@atarack/staff/sargun)
11:01:13 Join Soap [0] (n=Soap@rockbox/staff/soap)
11:01:13 Join tedrock [0] (
11:01:13 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
11:01:13 Join SirFunk [0] (
11:01:13 Join Galois [0] (
11:01:13 Join hax [0] (n=hax@httpcraft/hax)
11:01:13 Join Ryuu [0] (n=JD@unaffiliated/helldragon)
11:01:13 Join Caliban [0] (
11:01:13 Join Entasis_ [0] (
11:01:13 Join zylche [0] (
11:01:13 Join markun [0] (n=markun@rockbox/developer/markun)
11:01:13 Join merbzt [0] (
11:01:13 Join scorche [0] (i=HydraIRC@rockbox/administrator/scorche)
11:01:13 Join gromit` [0] (
11:01:13 Join Mouser_X [0] (
11:01:13 Join crashmatrix [0] (
11:01:13 Join andrew__2 [0] (
11:01:13 Join chrisjs169_ [0] (
11:01:13 Join alberink [0] (
11:01:13 Join RaRe [0] (
11:01:13 Join joshin [0] (n=joshin@unaffiliated/joshin)
11:01:13 Join blithe [0] (
11:01:13 Join DataGhost [0] (n=dataghos@
11:01:13 Join FOAD [0] (
11:01:13 Join gtkspert [0] (
11:01:13 Join bospaadje [0] (
11:01:13 Join amigan [0] (i=dcp1990@unaffiliated/amigan)
11:01:13 Join Weiss [0] (
11:01:13 Join crashd_ [0] (
11:01:13 Join Seed [0] (
11:01:13 Join Lynx_ [0] (
11:01:13 Join Neovanglist [0] (
11:01:13 Join Anakin [0] (i=0@unaffiliated/anakin)
11:01:13 Join Farp [0] (i=Farp@
11:01:13 Join qwm [0] (
11:01:13 Join GodEater [0] (
11:01:13 Join preglow [0] (n=thomj@rockbox/developer/preglow)
11:01:13 Join dionoea [0] (
11:01:13 Join Nimdae [0] (
11:01:13 Join Kohlrabi [0] (
11:01:13 Join DiDjCodt [0] (
11:01:13 Join guyzmo [0] (
11:01:13 Join Presence [0] (
11:01:13 Join pixie [0] (
11:01:13 Join sslashes [0] (i=sslashes@
11:01:13 Join feisar [0] (
11:01:13 Join Hadaka [0] (
11:01:13 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
11:01:13 Join toer [0] (
11:01:13 Join luckz [0] (
11:01:13 Join ze [0] (
11:01:13 Join Ave [0] (
11:01:13 Join bagawk [0] (n=lee@unaffiliated/bagawk)
11:01:13 Join krp [0] (
11:01:13 Join prashantv [0] (i=prashant@wagner.orchestra.cse.unsw.EDU.AU)
11:01:13 Join Battousai [0] (n=bryan@gentoo/developer/battousai)
11:01:13 Join Bjoern-Erik [0] (
11:01:13 Join h4ngedm4n [0] (
11:01:13 Join godzirra [0] (
11:01:13 Join lostlogic [0] (n=lostlogi@rockbox/developer/lostlogic)
11:01:13 Join [Ray] [0] (
11:01:13 Join Overand [0] (i=overand@pdpc/supporter/active/Overand)
11:01:13 Join enyc [0] (
11:01:13 Join [mbm] [0] (i=mbm@openwrt/developer/mbm)
11:01:13 Join crwl [0] (
11:01:13 Join sbeh [0] (
11:01:13 Join J [0] (
11:01:13 Join w0rd54 [0] (
11:01:13 Join idnar [0] (i=mithrand@unaffiliated/idnar)
11:01:13 Join daurn [0] (i=daurn@unaffiliated/daurnimator)
11:01:13 Join SaLoMoN [0] (
11:01:13 Join gregj [0] (
11:01:13 Join lost|X40 [0] (
11:01:13 Join sneakums [0] (
11:01:13 Join mtoups [0] (n=mtoups@HENSON.ISR.CS.CMU.EDU)
11:01:13 Join DogBoy [0] (n=john@unaffiliated/dogboy)
11:01:13 Join GodEater_ [0] (
11:05:23amiconnAh, hi logbot_ ;)
11:05:32amiconnThis probably means the site is also back
11:06:07lazkaamiconn: - and btw. i can abx some 200kb MP3s using my crappy PC loudspeakers ;)
11:06:52peturmarkun: ping
11:07:40amiconn*** ping error: markun: black hole detected
11:08:13peturhmmmm amiconn: bit early to drink, no? :p
11:08:54 Join ender` [0] (
11:10:21 Part Llorean
11:11:54 Join pixelma [0] (i=57b9f63b@rockbox/staff/pixelma)
11:13:35 Quit Rob2222 (Read error: 110 (Connection timed out))
11:17:22peturanybody from the Netherlands here that goes to concerts there?
11:17:40 Join stripwax [0] (i=5607d648@gateway/web/cgi-irc/
11:18:33 Quit ^BeN^ ("MULEz SCRIPT: Don't Take Life Seriously, It Is Not Permanent.")
11:19:40 Quit Redbreva ("ChatZilla [Firefox]")
11:22:11 Quit stripwax (Client Quit)
11:23:13*petur drives off to go see/hear bjork and hopefully tape too if they don't do bodysearches...
11:23:34scorchecavity searches?
11:24:04peturI wouldn't want to hide my h380 that way :/
11:24:14 Join ^BeN^ [0] (n=Paprica@
11:24:29 Quit petur ("on to recording field test :)")
11:29:05 Quit pixelma ("CGI:IRC (EOF)")
11:29:42 Join pixelma [0] (i=57b9f63b@rockbox/staff/pixelma)
11:34:08 Join hannesd [0] (
11:37:46 Quit amiconn (" HydraIRC -> <- Po-ta-to, boil em, mash em, stick em in a stew.")
11:38:37 Join obo [0] (
11:40:10 Quit Guile`` ("KVIrc 3.2.0 'Realia'")
11:40:35 Quit dan_a ("using sirc version 2.211+KSIRC/1.3.12")
11:44:46 Join dan_a [0] (n=dan_a@
11:46:58 Quit The-Compiler ("Connection reseted by god")
11:52:25 Join printfXh4 [0] (
11:53:54 Join My_Sic [0] (
11:57:20 Quit pixelma ("CGI:IRC 0.5.7 (2005/06/19)")
11:58:57 Join z35 [0] (
11:59:28 Join noobie [0] (i=dad7437b@gateway/web/cgi-irc/
11:59:46noobiehi guys
12:00:16noobieMy ipod mini freezes when ever i plug USB to computer
12:00:24noobieand wont mount
12:00:38 Part maffe
12:00:49 Join The-Compiler [0] (
12:00:52 Join maffe [0] (
12:03:31bluebrothernoobie: then reset it and use emergency disk mode
12:04:02bluebrotherthere is a known problem with some minis not rebooting into disk mode properly
12:04:37noobiehow do i boot into disk mode?
12:04:49noobieand do i copy yhe songs over through that mode?
12:04:52bluebrotherreset the ipod, then hold Play + Select
12:05:37bluebrotherthe player behaves as external drive in that mode. Much similar to usual disk mode
12:05:55bluebrotherbut it won't reboot if you disconnect usb, you'll need to reboot it manually.
12:06:38noobiehaving a play now thanks
12:07:06noobiewhere do i put the mp3? what directory?
12:07:36bluebrotherdepends on you
12:15:46 Quit z35 (Read error: 110 (Connection timed out))
12:17:51 Join lee-qid [0] (
12:19:56noobiehow do i turn it on?
12:20:05noobiei turned it off by holdng pause
12:21:19noobieo menu + enter
12:21:56bluebrotherplease check the manual ...
12:22:29noobieok one more Qs :P can u give me the link to themes
12:22:42bluebrothercheck the wiki
12:22:51bluebrotheror the manual ...
12:22:54bdgrauenoobie: extras on the rockbox website
12:23:06noobiethanks for ue help guys
12:23:26noobiewill be 100% apple free soon :D
12:23:29 Quit noobie ("CGI:IRC")
12:25:15*bluebrother sighs
12:27:49lazkai just tried the sansa simulator binary on linux using wine. everything seems to work, even sound. shouldn't this be mentioned in the wiki??
12:28:23bluebrotherfeel free to add it, but when running linux you most likely want to build the sim yourself
12:29:08bluebrotheralso, the main purpose of the sim is debugging, so running in wine somewhat defeats this purpose ...
12:41:34 Join Nico_P [0] (
12:43:17Nico_PJdGordon: hi
12:46:18 Quit perplexity (Read error: 110 (Connection timed out))
12:50:08 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
12:51:29 Join kubiix [0] (
12:52:59 Join n1s [0] (
12:53:23JdGordonhey Nico_P
12:53:44JdGordonsorry about the server, my mum turned the computer off last night and i didnt notice till a few hours ago
12:54:27Nico_Pno problem ;)
12:55:00Nico_Phave you had time to look at my changes ?
12:55:03 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
12:55:49safetydanAm I wrong in thinking the emac on coldfire isn't useful if you do a lot of 16x16 ->32 operations?
13:01:09***Saving seen data "./dancer.seen"
13:01:34 Join JdGordon_ [0] (n=jonno@rockbox/developer/JdGordon)
13:02:43 Join amiconn [0] (
13:03:16XavierGrI am glad to say that after saratoga's yesterday commits wma is real time on my H300 (128 wma track). It is struggling but it didn't had a single gap :)
13:04:21amiconnI got 122% realtime with 128kbps wma on H180
13:04:39XavierGrI am testing with test_codec now
13:06:07amiconnNow compare this with 542% realtime for mp3 (~200kbps) on coldfire...
13:06:30XavierGrlol I got exactly 122.22% on my H300 too amiconn
13:06:35 Join Guile`` [0] (n=Guile@
13:07:24amiconnOn arm wma might now even be faster than mp3... will test in a minute
13:08:49XavierGrwell saratoga was mumbling something about not utilizing iram efficiently and putting the wrong data in it, so hopefully coldfire will become fast as hell too
13:09:27amiconnI think this will only be possible with making real use of the EMAC
13:09:51*safetydan attempts to direct peoples optimisation efforts towards Speex :)
13:09:54amiconnaac is also roughly on par with mp3 on arm, but dog slow on cf
13:14:36Nico_Pamiconn: I'll test the patch you gave me yesterdat
13:14:45 Quit ompaul (Read error: 110 (Connection timed out))
13:15:27*amiconn concludes that Nico_P uses a qwerty keyboard ;)
13:15:54Nico_Phehe... azerty actually (french)
13:16:09*amiconn uses qwertz (german)
13:16:55 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
13:18:18 Quit JdGordon (Read error: 110 (Connection timed out))
13:18:19amiconn178% realtime on ipod (128kbps wma)
13:21:47 Quit JdGordon_ (Read error: 110 (Connection timed out))
13:23:36 Quit TiMiD[FD] ("leaving")
13:24:22Nico_Pamiconn: I'm sorry to report still no crash
13:24:43Nico_Pwith incorrect results that show there were alignment proiblems
13:25:23 Join spiorf_ [0] (
13:25:29*Nico_P is off for lunch
13:25:41n1samiconn: do you think FS #5264 can be closed?
13:25:50 Join DerPapst [0] (
13:26:57n1shaha, that is one of the three tasks blocking a 3.1 release :-) (according to the tracker)
13:28:42amiconnI don't know. No further replies for omne month. But all that happened before my fixes
13:29:00 Join thegeek [0] (
13:29:35n1swell, I'll close it then and they will probably post a new bug if the issue persists (that task has way to many comments anyway...)
13:31:03amiconnThe tracker doesn't know about the biggest tast that's blocking a 3.1 release....
13:31:30n1sgood work on fixing the crashes btw :-)
13:31:38bluebrotherthe release of 3.0 first?
13:31:54n1sI wonder if anyone uses the target release stuff though
13:32:42amiconnn1s: Thanks
13:32:47bluebrotherwell, shouldn't we start using it? That roadmap feature of flyspray is really nice.
13:33:18amiconnThe clock setup needs a bit more cleanup, and I hope to be able to help dan_a with some PP5002 fixes as well
13:33:33n1sbluebrother: but will we ever release? we may be giving someone false hopes :-P
13:33:40amiconnAnd I'll go 80MHz soon...
13:34:18bluebrothern1s: well, users might actually look at the roadmap task list and think we're that −−> <−− close to a release ...
13:34:48 Quit ^BeN^ (Read error: 110 (Connection timed out))
13:34:52amiconnI'm just not sure yet whether it's better to use precalculated register values for each frequency setting, or implement value calcuulation similar to what the OF does
13:37:55 Quit spiorf (Read error: 110 (Connection timed out))
13:38:42 Quit bdgraue (Read error: 113 (No route to host))
13:39:19n1samiconn: I found out that using unsigned ints for some stuff (for example width and height in the screen struct) saves a fair amount of code size. It is tricky buissnes though. (maybe not for those values but some other stuff...)
13:40:13*bluebrother notices his h120 showing some arbitrary time without the RTC mod installed
13:43:05 Join moos [0] (
13:49:40Nico_Pbluebrother: I totally agree on the roadmap (it's roughly what gregj was saying we should do :p )
13:50:03Nico_Pamiconn: what's "the biggest task that's blocking a 3.1 release"
13:50:07bluebrotherwell, then just start using that ;-)
13:50:09Nico_Pviewports ?
13:50:18 Join Wiwie [0] (
13:50:23amiconnThe 3.0 release....
13:50:28bluebrotherhow about a "flyspray cleanup week"? (or month)?
13:50:50Nico_Pbluebrother: that's needed too
13:50:54bluebrotherwe could ask users to verify open bugs ...
13:51:18bluebrotherand close feature request that are highly unlikely to happen as "later"
13:51:25 Join Crash91 [0] (n=crash@
13:51:49safetydanhrmm, ICONST_ATTRing all static tables in speex only gains 9% on coldfire
13:53:58n1ssafetydan: have you tried putting any working buffers in iram?
13:54:21safetydann1s, at the moment I'm just trying very simple things
13:54:32 Join perplexity [0] (
13:55:05n1ssafetydan: iirc putting those buffers in iram have given the larest speedups...
13:55:30safetydann1s, I'll take a look at that then.
13:55:38safetydanbut for the moment it's sleep for me
13:55:39 Quit safetydan ("Ex-Chat")
13:58:05 Join Xerion [0] (
13:59:35 Quit DerPapst (Read error: 104 (Connection reset by peer))
14:05:39 Quit The-Compiler (Read error: 104 (Connection reset by peer))
14:06:09 Join Thundercloud [0] (
14:07:41 Nick Crash91 is now known as Defender (n=crash@
14:08:20 Nick Defender is now known as xDefender (n=crash@
14:08:23 Part xDefender
14:24:44 Join Arathis [0] (
14:25:55 Join z35 [0] (
14:30:16 Join chrisjs169 [0] (
14:33:52 Join Jaymuhz [0] (
14:33:56 Part Jaymuhz ("Leaving")
14:36:47 Join Crash91 [0] (n=crash@
14:38:16n1sbluebrother: why did you use a build with RTC enabled?
14:38:58bluebrotherbecause I want to do the mod ;-)
14:39:55bluebrotherfunny thing is that using rambox it works fine, rombox not.
14:42:07XavierGrbluebrother: did you find the RTC chip that roolku suggests?
14:42:21bluebrotheryes, I have both of them now
14:42:45 Nick Soap__ is now known as Soap_ (
14:51:10XavierGrbluebrother: where did you find them? ordered them from the company?
14:54:12 Join ^BeN^ [0] (n=Paprica@
14:54:53 Quit Crash91 (Remote closed the connection)
14:56:25 Join xDefender [0] (n=BotsUser@
14:57:12 Join Crash91 [0] (n=crash@
14:57:13xDefenderWelcome Crash!!
15:01:13***Saving seen data "./dancer.seen"
15:03:00xDefenderSun Jul 8 16:02:29 2007
15:03:35 Join DerPapst [0] (
15:04:41Crash91does anyone know how close we are to getting USB on Sansa? (Can you estimate a date?)
15:04:49Crash91lol :)
15:09:55 Join Soul-Slayer [0] (n=Administ@
15:11:03Soul-SlayerHey, does anyone elses gigabeat occasionally display 'Battery empty' when they first turn it on, then when they try again it'll boot into Rockbox fine? Or am I alone in that :$?
15:11:08 Quit Nico_P (Remote closed the connection)
15:11:47n1sSoul-Slayer: does rockbox say "Battery empty" or the gigabeat loader?
15:11:55Soul-SlayerGigabeat Loader
15:12:59 Quit miepchen^schlaf (Read error: 113 (No route to host))
15:13:24 Join miepchen^schlaf [0] (
15:13:45Soul-SlayerI have a feeling my battery is shot anyway, yesterday I fully charged it and after 2 hours it had only lost about 6%, then it suddenly jumped down to 70% remaining. But I'm not sure why it'll display the battery as empty, then when I try again it realises it's not
15:14:19 Join Nico_P [0] (
15:17:47xDefenderSun Jul 8 16:17:16 2007
15:18:18 Join webguest60 [0] (i=864c0358@gateway/web/cgi-irc/
15:19:06 Nick xDefender is now known as Crash91_ (n=BotsUser@
15:19:12webguest60Crash91: could you use another channel to play with your bot?
15:19:46Crash91oh sorry...i...i dunno why the last message came here
15:19:53Crash91i am*
15:20:52 Quit webguest60 (Client Quit)
15:21:30Nico_PjhMikeS: what happends if you use queue_reply on a message that wasn't sent using queue_send ?
15:23:22jhMikeSNico_P: nothing...queue_reply simply replies to the last message dequeued that was sent using queue_send. calling it multiple times on the same message only replies once as well.
15:25:27Nico_PjhMikeS: would you have a bit of time to have a look at my code ?
15:27:57Nico_Pit's in JdG's svn: svn://
15:28:05Nico_Pyou'll want to look at testplugin.c
15:29:32Nico_Psome parts of the code are leftovers from the old test plugin where there was no threading (the body of playback_thread() and what follows the queue_reply() call in the main plugin func)
15:29:40 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
15:29:58GodEater_I think the battery in my iPod is hosed.
15:31:09GodEater_which is a little disappointing given it's less than a year old
15:31:29jhMikeSNico_P: I've gotta do a co of that?
15:31:55Nico_PjhMikeS: yes. it'll be very quick, there aren't a lot of files
15:32:11 Quit z35 (Read error: 110 (Connection timed out))
15:32:48DerPapstGodEater_: it's not even 3/4 year old
15:33:28jhMikeSyeah, only 10 :)
15:33:50DerPapstGodEater_: the battery of mine is still ok.. used frequently.
15:34:59jhMikeSNico_P: Any part in particular?
15:36:27Nico_PjhMikeS: maybe the beginnings of the threading logic
15:37:21Nico_Pin particular, what do you think of passing the arg data through structs like that ?
15:37:37jhMikeSthe if (result) queue_reply() bit isn't right. queue_reply really should just be called always for any message queue_send could send
15:38:10jhMikeSthat's what it's for. recording passes filenames that way with queue_send
15:38:52Nico_PI corrected that
15:39:03 Join Kreative_Think_R [0] (i=485c0f23@gateway/web/cgi-irc/
15:39:29Kreative_Think_RNe New For Rockbox On Sansa ?
15:39:49jhMikeSthere's really no reason to be choosy about calling queue_reply since it returns very fast if not in a send state. you could enable the queue_in_queue_send function which takes fewer params though.
15:40:12 Quit atsea- (Read error: 104 (Connection reset by peer))
15:40:21Nico_Pwhat's that ?
15:40:48Kreative_Think_Rcan rockbox play divx knw ?
15:41:03jhMikeSNico_P: asking me?
15:41:20Nico_PjhMikeS: yes, what's queue_in_queue_send ?
15:41:57Crash91Kreative_Think_R: No! Lets get mpegplayer up properly first then think about DivX
15:42:17jhMikeSqueue_in_queue_send returns true if the last message dequeued was sent by queue_send and the wait waiting thread wasn't bumped. basically it's the checking part of queue_reply without the reply part
15:42:18*amiconn starts to understand the PP5002 clock setup :)
15:42:18 Quit Kreative_Think_R (Client Quit)
15:42:33n1swhat's a good way to check for all archos targets? just #ifdef CPU_SH ?
15:42:35Crash91Kreative_Think_R: Sandisk has also contacted DivX to help them use DivX on the Orig Firmware
15:43:18Nico_PjhMikeS: ok, but calling queue_reply each time is all right, isn't it ?
15:43:30Soul-SlayerDoes anyone know where it would be possible to get a replacement Gigabeat F20 battery?
15:43:40jhMikeSit's fine and intended to be used like that
15:43:55 Join atsea- [0] (i=ariel@gateway/tor/x-bd4200a702e2811c)
15:44:22GodEater_DerPapst: good to hear yours is okay - there's definitely something wrong with mine though
15:44:42GodEater_I used an AC wall charger with it on Friday night - then turned it off once it was fully charged
15:44:57GodEater_turned it on this afternoon and got the "Very Low Battery" warning
15:45:04GodEater_that can't be healthy
15:45:21DerPapstno it can't
15:45:34DerPapstluckily you're still under warranty
15:45:48GodEater_but I've ordered a new battery for it anyway
15:45:52GodEater_an 850maH
15:46:01GodEater_instead of the stock 600maH
15:47:20jhMikeSNico_P: I don't see where you allocated the struct queue_sender_list and called queue_enable_queue_send
15:47:28Crash91GodEater_: What type is the battery?Li-on?
15:47:42GodEater_Crash91: yes
15:48:06Crash91hmm: how long have you been using it?
15:48:18GodEater_since I bought it...
15:48:33Crash91which was when.... >_>
15:48:38Nico_PjhMikeS: the struct queue_sender_list is declared before playback_thread()... I don't think I called queue_enable_queue_send
15:48:40GodEater_Sept. 06
15:48:45DerPapsthopefully it doesn't ballon as petur'S did
15:49:14GodEater_fingers crossed
15:49:32GodEater_although if it did
15:49:37GodEater_the case should just pop off
15:49:46Crash91hmm....thats bad...but li-on batteries have very long lives unless you keep them plugged in for excessive amounts of time
15:49:48GodEater_it's not screwed together like an H140
15:50:02jhMikeSah oops
15:50:39DerPapstbut imagine it goes one step further than just balloning ;)
15:50:51GodEater_I'd rather not!
15:51:12DerPapsthow much was it?
15:52:41Nico_PjhMikeS: is there a particular reason why the queue_send and related functions aren't in the plugin API ?
15:52:50DerPapstsounds ok...
15:53:19jhMikeSNico_P: noone has used them there yet?
15:53:28Nico_Pok :)
15:53:55*amiconn wonders what currency '?' is
15:54:50 Quit perplexity (Read error: 110 (Connection timed out))
15:55:11DerPapstif that is spelled correctly :P
15:56:23 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
15:56:36 Nick Crash91 is now known as cras2 (n=crash@
15:56:58 Quit amiconn (Nick collision from services.)
15:56:59 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
15:57:04GodEater_dodgy irc client again amiconn ? :)
15:57:09GodEater_it was indeed GBP
15:57:18amiconnWLAN disconnected for some reason
15:57:19Soul-SlayerIs it really impossible to find a replacement gigabeat F20 battery o.0? Can't find anything other than an 850 mah one and apparently the stock one is 2000mah
15:57:44*amiconn prefers using the standard 3-letter abbreviations for currencies
15:57:44GodEater_Soul-Slayer: where did you read the stock one is 2000mAh ?
15:58:01 Nick cras2 is now known as crash91 (n=crash@
15:58:22Soul-SlayerWell, in Rockbox, under 'Battery capacity'
15:58:26Soul-Slayerit defaults to 2000
15:58:32Soul-SlayerAnd doesn't even go as low as 850
15:58:35 Quit printfXh4 (Read error: 113 (No route to host))
15:58:53 Part chrisjs169
15:59:07GodEater_Soul-Slayer: that default means nothing
15:59:11Soul-SlayerOh... Right =$
15:59:16 Join perplexity [0] (
15:59:26Soul-SlayerSo this replacement is infact larger... Great ^^
15:59:44Soul-SlayerThen why does it say 2000 within Rockbox? What does that mean?
16:00:08GodEater_it means nothing because as far as I'm aware, the gigabeat hasn't been calibrated yet
16:00:13*DerPapst wonders when they introduce batteries based on fullerenes
16:00:19Soul-SlayerIt hasn't? That explains SO much
16:00:37Soul-SlayerI was wondering why my battery meter went 100, 99, 98, 97, 96, 70...
16:00:41DerPapstno memmory effect, larger capacities
16:00:55GodEater_DerPapst: how about fuel cells ?
16:01:53 Quit jhulst (Read error: 104 (Connection reset by peer))
16:03:55 Join Ben1422 [0] (i=3aa17116@gateway/web/cgi-irc/
16:04:26crash91GodEater_:Why fuel cells? i will not settle for anything below a fusion reactor
16:05:03DerPapstmatter, anti-matter reactor
16:05:57*DerPapst has watched lots of star trek episodes :P
16:05:58crash91knew there was something better :p
16:06:22*crash91 agrees with DerPapst,
16:06:29 Join Ben1141 [0] (i=3aa17116@gateway/web/cgi-irc/
16:06:43Ben1141Where is a sample video?
16:06:51Ben1141(to try)
16:06:58Ben1141Any links?
16:07:02DerPapston the pliginmpegplayer wiki
16:07:35 Join My_Sic [0] (
16:07:37DerPapstbut there is that neat serchbox
16:07:46DerPapsttypein mpeg player
16:07:50Ben1141OH OK
16:07:50DerPapstand you'll find it
16:08:54crash91DerPapst: I Just Can’t Do It Captain
16:08:57 Quit Ben1422 ("CGI:IRC (Ping timeout)")
16:09:16DerPapstheh ;)
16:09:24 Quit moos ("Glory to Rockbox")
16:09:27 Join Nibbier [0] (
16:09:27lazkaWhat is the new LCD sleep mode on the sansa?
16:10:35crash91im guessing its something like the LCD patch
16:10:37Ben1141 4:3 is widecreen? Where 16:9 is square?
16:10:51crash9116:9 = widescreen
16:11:07crash914:3 = square-ish (normal TVs)
16:11:24Ben1141Okay. Great. Thanks.
16:11:27crash91but as you can see, its not square as it is Four by Threee
16:11:42crash91and why doesnt threee have threee e's?
16:14:18 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
16:15:09jhMikeSLCD sleep mode actually unpowers the chip. It's an option because it can put a big delay in having the backlight come on again.
16:15:54 Quit Ben1141 ("CGI:IRC (EOF)")
16:16:03lazkajhMikeS: ahh, thanks
16:17:29lazkajhMikeS: do you think it will save much battery?
16:22:11jhMikeSI've no idea really. I guess test if it matters or not.
16:22:38crash91yeah...why does the original firmware last 16 hours while rockbox lasts like 4 -5 ?
16:23:33 Join My_Sic [0] (
16:23:34jhMikeSon e200 I thought it ran 12 in battery tests. That may improve shortly with some clocking work.
16:23:39 Join Davide-NYC [0] (
16:24:23crash91is CPU scaling enabled? i always see boost=100% when i go into Debug>audio thread
16:24:34crash91even while completely idle
16:24:43jhMikeSon e200 is isn't yet
16:25:05crash91Scaling should up life by at least 1 hour
16:25:54Davide-NYCgood morning all (at least morning for me)
16:26:03jhMikeSuse of the COP for audio decoding should help bring the boost down too
16:26:09crash91Good evening here
16:26:25Davide-NYCI cannot find any good documentation (no matter how brief) on the new test_codec plugin
16:26:40jhMikeSDavide-NYC: hey! what the? you're actually in here?
16:26:51Davide-NYCSunday morning
16:27:03Davide-NYCbeen working (read $$)
16:27:18Davide-NYCthat's why I am a total no show for the last 3 or 4 months
16:27:20*jhMikeS wants some of those $$ :)
16:27:32Davide-NYCso do I (debt)
16:28:01Davide-NYCanyhoo, what's all of this about WMA and test_codec... sounds yummy
16:28:46 Join Explosivo [0] (
16:29:10Davide-NYCI'd like to merge this into this and settle on a copyright free test file as a standard practice.
16:29:48Davide-NYCthat way folks can have a standard place to put test results and the test_codec plugin should make it pretty easy
16:30:19*jhMikeS offered his own music and got no reply (not even interest to hear anything) :P
16:31:00Davide-NYCnot by me bro
16:31:01*crash91 thinks everyone should join #crash
16:31:02bluebrotherjhMikeS: what type of music is it?
16:31:49jhMikeSjust stuff I did with MIDI on a Soundblaster
16:32:13*Davide-NYC ducks
16:32:15*crash91 sees why no one wants to listen to it...
16:32:26jhMikeScan't say I know what style you'd call it
16:33:36jhMikeSbut I'm just putting it on the table and offering it as a standard
16:36:25Davide-NYCthe fishbone track I chose lacks dynamic range. It's also a bit noisy. The style of music doesn't matter really. As long as it get's the job done. The procedure is pretty well defined already in the wiki, it just needs to be updated to include the new test_codec plugin to make it easier.
16:37:49 Quit Crash91_ ("received SIGTERM")
16:38:34 Join xDefender [0] (n=BotsUser@
16:40:18lazkaAm I the only one noticing sluggish scrolling while playing music lately?? I don't remember it this bad.. :/
16:42:00 Quit xDefender (Client Quit)
16:42:15 Join Crash91_ [0] (n=BotsUser@
16:42:26jhMikeSlazka: no...not the only one. are you inclined to find when it started? I'm sure it's probably my fault. ;)
16:43:25jhMikeSit seems to start boosting the codec thread a bit sooner than it did :\
16:49:05lazkabtw. is there now a known way to avoid the database refresh of the original firmware?
16:50:53crash91unless youre in mtp
16:51:13crash91lazka: in the future, speicfy youre talking about a sansa :)
16:52:02lazkacrash91: sry, i always forget that
16:58:09Davide-NYCquestion: why was test_codec removed from SOURCES?
17:01:17***Saving seen data "./dancer.seen"
17:03:16 Quit hannesd ("Client suicide")
17:03:40 Quit Explosivo ("ChatZilla [Firefox]")
17:07:32n1sDavide-NYC: it's a testing plugin for developers, not intended for users, just like test_fps
17:07:35*crash91 wonders why noone cares about #crash
17:09:33Davide-NYCn1s: so I just add a test_codec line to apps/plugins/SOURCES and recompile?
17:09:44n1sDavide-NYC: yes
17:10:15Davide-NYCI'm getting some errors in the h1x0 sim built under cygwin
17:10:35n1sDavide-NYC: then long-press select on a file and open with test_codec
17:11:17Davide-NYCnevermind about the sim error
17:11:27Davide-NYCforgot to make install (doh!)
17:16:30lazkaok, forget it... scrolling has alway been this sluggish..
17:20:04crash91Davide-NYC: join #crash again sometime!
17:20:09 Quit perplexity (Read error: 110 (Connection timed out))
17:20:35*Davide-NYC waits for the puchline
17:20:49*Davide-NYC *punchline
17:21:44*crash91 feels sorry for Davide-NYC
17:21:59Nico_PjhMikeS: about my buffering code, what do you think about the use of structs to pass argument data through the queue_* calls ?
17:22:20crash91***Crash wonders why noone cares about #crash
17:23:02crash91and why are there no ops on this channel?!>!
17:23:56n1sCrash91: please stop spamming, and the ops are here but they don't have their op hats on all the time :-)
17:24:31crash91and in what way am i spamming :p
17:25:11*Davide-NYC hugs ignore
17:25:14jhMikeSNico_P: didn't you ask that already?
17:25:38Nico_PjhMikeS: did you answer ? sorry if you have
17:25:54jhMikeSyes, that's another good use of it and recording does that
17:26:01crashd_not so bad i guess
17:26:04crashd_not looking forward to another week
17:26:20 Quit crash91 (Client Quit)
17:27:18Nico_PjhMikeS: I thought that part was about queue_send :)
17:27:26 Join The-Compiler [0] (
17:30:07*jhMikeS is examinig more closely
17:31:12jhMikeSthat'll work as long as another thread can't come along and clobber the data
17:34:38Davide-NYCSuggestion for the test_codec plugin: the log file should be delimited in such a way that a person could cut'n'paste the output to the wiki and get a reasonably formatted table there
17:35:12Davide-NYCmaybe using the | symbol instead of LFs within a particular result.
17:35:24Davide-NYCthen LF between file results
17:37:28 Join mattzz [0] (
17:46:17jhMikeSNico_P: this new buffer is going to send a message and wait every time a codec wants some data?
17:48:28Nico_PjhMikeS: what do you mean by "send a message and wait" ?
17:50:38amiconnjhMikeS: Do we really need invert on colour targets?
17:50:56jhMikeSit appears that a message is sent to get any piece of data at all instead of just reading it on the calling thread unless I've just not gotten familiar enough with it
17:51:29jhMikeSamiconn: If not then maybe it should go on all of them since it's not the only one with that
17:51:52jhMikeSiCatcher looks better inverted though :)
17:51:54amiconnI don't know of any. Only b&w and greyscale targets
17:52:10amiconnX5 also has that nasty sleep option
17:52:27amiconnThat should go away, and display sleep always enabled
17:53:13*amiconn almost never uses his X5
17:53:13Nico_PjhMikeS: the calling thread calls bufgetdata to know how much data he can read
17:53:17jhMikeSsome find the extra delay to restart the lcd from standby annoying though...myself included. it's ok if I'm not going to touch it for awhile.
17:53:56amiconnThen make it sleep a fixed amount of time after the backlight goes off
17:54:02amiconn(e.g. 5 seconds)
17:54:48jhMikeSNico_P: but why is a message send/post needed to get that information? that kind of task switching can be slow.
17:54:53amiconnUsually I am easily annoyed by delays, but on X5 I couldn't even notice the difference between lcd sleep enabled and disabled
17:55:02jhMikeSmmm...I ususually use around 30s or more.
17:55:22Davide-NYCdoes anyone know of a command line WMA encoder?
17:55:29Nico_PjhMikeS: it's true that a message probably isn't needed
17:56:20 Join crash91 [0] (n=crash@
17:56:21Crash91_Welcome Crash!!
17:56:54Nico_PjhMikeS: that's why I asked you to look at the code :) I have no experience in threaded programming so review is good
17:57:10 Quit crash91 (Client Quit)
17:57:59jhMikeSamiconn: I'm quite aware of the sluggishness of bringing the display back since the delays needed are much longer
17:58:27Nico_PjhMikeS: the thing is the buffer might have to be filled in order for the required amount of data to be available
17:59:46jhMikeSNico_P: so, if the data isn't available, then block. otherwise let the caller update the read position. some mutexing in the correct places is going to be way more efficient.
18:01:11Nico_PjhMikeS: I'm starting to wonder if a buffering thread is really necessary...
18:02:15 Join PaulJam [0] (
18:02:43jhMikeSit might not be nescessary but it might be nice. the thing with threaded programming is to lock things together as little as possible or you lose the benefit.
18:05:06Nico_Phmm, I think I'm losing sight of what it should do. in fact it should almost only check wether the buffer has enough data and fill it when appropriate, shouldn't it ?
18:05:47 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
18:07:07jhMikeSit could block if the request can't be satisfied or even be pollable without a block or both
18:07:59jhMikeSrecording performs no locking at all but just polls the status of the data (but that is much simpler overall)
18:08:30Davide-NYCafternoon y'all
18:08:34 Quit Davide-NYC ("ChatZilla [Firefox]")
18:09:34Nico_PjhMikeS: I'll look at the recording code, it might prove inspirational :) what do you mean by "be pollable without a block or both"
18:09:45jhMikeSif the COP runs the codec and the buffer runs on the CPU for pp targets with practically no locking, that's the best situation since the cop can just whiz through and get the data
18:10:16 Join pondlife [0] (
18:10:17jhMikeSthe call for data could optionally block until the data is ready or just return with a status of some sort
18:11:05Nico_Pok, so in fact there's no need for the message queue ?
18:11:25Nico_Pat least for the Q_BUF* messages ?
18:11:45 Join miepchen^schlaf [0] (
18:13:35jhMikeSnot strictly unless a wait is needed/wanted for the data or some reply is needed from the thread
18:15:01 Join z35 [0] (
18:15:18jhMikeSone advantage to polling is letting the codec free to keep checking its own status like whether to quit or not and that can avoid deadlocks.
18:16:46pondlifeNico_P: I was vaguely thinking that any buffering thread would only be for filling the buffer, not necessarily for handling the API requests, if that's any help.
18:17:21pondlifeSorry I've not had time to look at your code since you started thread stuff.
18:17:22Nico_Ppondlife: that's what I was starting to think too
18:17:50Nico_Ppondlife: I think I'm just going to throw away most of the threading stuff I did, anyway :p
18:19:03 Quit Crash91_ (Read error: 110 (Connection timed out))
18:19:05jhMikeSI'm thinking that codecs having a message queue would be beneficial since things like requesting a stop or a seek (relatively rare event) could be done simply and securely.
18:19:23Nico_PjhMikeS: I think i'll go with polling... just to make sure we're talking about the same thing, it means that (for example) bufgetdata will return an error status if it can't provide enough data ?
18:20:05 Quit bluebrother (Read error: 113 (No route to host))
18:20:15Nico_P... instead of blocking to try to fill the buffer in order to be able to provide the data ?
18:20:17pondlifeNico_P: I think bufread/bufgetdata need to block... otherwise what will the calling routine do if there's not enough data...
18:20:42pondlifeIt would have to wait I think - hence blocking but in a different place.
18:20:51Nico_Pthat's true
18:20:54jhMikeSright, and it can do other stuff and check it again until the data is ready
18:21:19pondlifeI don't know if the codec callbacks allow that flexibility.
18:21:52 Join bluebrother [0] (
18:22:06jhMikeSso, if the data isn't a message and return to the codec. the codec checks it's queue status and if nothings there to process, it checks the buffer status again
18:22:06pondlifeWhen a codec needs data, it needs data, right?
18:22:29pondlifeI thought the codec made the call to request more data?
18:22:47Nico_PI think it does
18:22:55jhMikeSand it would but the codec doesn't have to block there
18:23:37pondlifeAh, there's a response on the codec queue...?
18:23:54jhMikeSit's already a deadlock possiblity because how do you tell a blocked codec to quit?
18:24:19pondlifeIs that what's causing the STOP while resuming lock?
18:24:21jhMikeSwe already have a resume deadlock in the works :)
18:24:42pondlifeHey, the more the merrier :p
18:24:45jhMikeSI fixed it but by relying on timing and avoiding a yield in one spot
18:25:25pondlifeNeeds some proper design really.
18:25:40*pondlife hints
18:25:43Nico_PI don't really feel competent for such design
18:25:48pondlifeMe neither.
18:26:49pondlifeIt does "feel" that the codec handling is a little overcomplicated and too dependent on other threads/timing.
18:27:14pondlifeBut that's way OT for buffering/MoB.
18:27:40 Quit zylche ("+++ ATH0--;")
18:29:03jhMikeSway OT? it seems OnT since the buffer deals with both the metadata and audio data
18:29:24pondlifeI didn't think the codec interface would need to change.
18:29:32 Join bdgraue [0] (
18:29:48jhMikeSpondlife: that's one problem with playback.c. no mutexing between the codec and buffer and since yields come and go, that just throws it off when it changes.
18:29:49pondlifeJust the mechanism by which we pass the data pointers.
18:30:58 Quit DerPapst ("So Long And Thanks For All The Fish!")
18:31:30jhMikeSI just thought of the codec queueing and polling for buffer data today. it could keep things responsive and deadlock free.
18:32:16pondlifeSounds good... perhaps you could explain more... (especially to Nico_P and JdGordon!)
18:32:22 Quit merbanan (Remote closed the connection)
18:32:25jhMikeSbuffering should be abstracted away from playback.c of course. that's very important to fixing it all.
18:32:36Nico_PI'd be happy to help improve the playback code
18:32:52Nico_P...beyond MoB of course
18:34:06 Join Robin0800 [0] (
18:34:20jhMikeSwith buffering abstracted enough wouldn't anything on the buffer be easier?
18:34:32 Join BHSPitLappy [0] (
18:35:34 Join FalscherProphet [0] (
18:35:53jhMikeSwell, talking about this has renewed my interest in true seeking support in mpegplayer...that's a bugger since two decoders are reading it at the same time and they aren't nescessarily in sync.
18:35:55 Part FalscherProphet
18:36:52pondlifempegplayer might be an easier place to develop/test the buffering API...
18:37:17pondlifeI've not looked at the code, but it must be clearer than playback.c.
18:37:30 Quit inversions (Read error: 113 (No route to host))
18:41:06jhMikeSessentially reversing everything from how it is now seems to be the way to a thread is still needed for background buffering though.
18:41:43pondlifeWill that work with the Archos playback too?
18:42:58jhMikeSI would think so. I think the MAS codec should just be seen as a codec. why does playback need to know if the decoder itself is hardware or software?
18:43:34pondlifeMAS playback doesn't use playback.c, so I assume it has separate buffering at the moment..
18:43:48pondlifeBut this should be unified.
18:43:53jhMikeSit's all completely duplicated
18:46:22jhMikeSI suppose it would get encoded data out of the buffer the same way as a software one but it just wouldn't make pcm buffer calls to do output but could call a hardware driver API instead
18:47:41amiconnThe difference with the MAS is that we (ideally) do not want the "codec" to copy the data into a separate buffer, but operate in-place
18:48:28amiconnHowever, if we want to support simple compressing codecs instead of just pcm wav/aiff and mp3 (e.g. adpcm), we would need a separate buffer
18:50:25jhMikeSI suppose simple access to a pointer is fine, if a "codec" wants to copy it can copy or just use it in place. Things don't nescessarily copy now anyway for SWCODEC.
18:51:08amiconnThe bitswap would need to run *in* the main data buffer
18:51:25amiconn...and it needs to remember which parts are already swapped
18:51:44amiconnThis could probably be handled in the metadata header of the track
18:51:54jhMikeShow is the bitswap performed
18:52:56jhMikeSevery 32-bit int needs the bits reversed?
18:53:50amiconnNo, every byte
18:55:49jhMikeStoo slow to just never buffer unswapped data?
18:55:55 Join ptw419 [0] (
18:56:29amiconnThe bitswap is asm optimised nowadays
18:56:45amiconnPerhaps we could swap on load, that'd need experimentation
18:57:44jhMikeSalso for pcm or just mp3 data?
18:57:57amiconnpcm and mp3
18:58:08amiconnThis is not due to the mas, but due to the SH1 SPI
18:58:21amiconnFor Ondio, I'm even tempted to add a "multimedia read" to the mmc driver
18:58:41amiconnThis would save the current double-swapping
18:59:22 Quit spiorf_ (Remote closed the connection)
18:59:35amiconnThe MMC also uses SPI, and the SH1 SPI uses the reverse bit order to what everything else uses
18:59:43jhMikeSone would think the SH could have implemented a bit order switch :\
18:59:52amiconnSH2 and higher have that
19:00:18 Quit mattzz ("Leaving")
19:01:20***Saving seen data "./dancer.seen"
19:03:53 Join kaaloo [0] (
19:04:32 Part kaaloo
19:19:51bluebrotherXavierGr: around? Missed your question earlier ...
19:21:11*jhMikeS wonders what could possibly have added so much size to e200 with an lcd commit :\ all of recording hardly added that much.
19:23:22The-CompilerCan I install a firmware-file on the iPod without using iPodWizard?
19:23:24 Join krazykit [0] (
19:23:50The-CompilerMaybe with ipodpatcher?
19:24:23bluebrotheryou can write binaries with ipodpatcher. Just try the −−help option ;-)
19:27:03Nico_PjhMikeS, pondlife: you might want to look at what I just committed
19:27:14Nico_P(in Jd's SVN)
19:27:48The-Compilerbluebrother: It doesn't work: [ERR] Input file too big for buffer
19:27:59The-CompilerWith ./ipodpatcher /dev/sdb -wfb ../firmware5.5.bin
19:28:19The-CompilerI would like to replace the original-apple-firmware with a modified one.
19:28:38bluebrotherthen dd it to the hidden partition.
19:28:45bluebrotherand reinstall the bootloader afterwards.
19:29:02bluebrotheror try −−write-partition :)
19:30:56The-CompilerThe hidden partition has also other images than the firmware, hasn't it?
19:31:32The-CompilerWell, I'll give it a try
19:31:40bluebrothernone that I know of. Disc mode is in flash afaik
19:31:49bluebrotheryou could just save the partition before ...
19:32:13pondlifeNico_P: Do you need line 751 - "read = MIN(read, GUARD_SIZE);" ?
19:34:30amiconnbluebrother: On the video there are other images than just the firmware
19:35:19bluebrotherah. Only on the video?
19:37:55 Join mario [0] (
19:38:08 Nick mario is now known as Covert (
19:38:43 Nick Covert is now known as Covert28 (
19:39:09 Join merbanan [0] (n=banan@
19:39:29Covert28Is there a ferrari theme?
19:40:23 Join spiorf [0] (
19:40:33Nico_Ppondlife: that's to read small chunks... it's not really necessary
19:40:52Nico_Pwithout it, all the availale data is read at once
19:41:19pondlifeCan bufgetdata return a larger value than GUARD_DATA?
19:41:50pondlifeI assumed not, but I suppose it could do.
19:41:58pondlifeGUARD_SIZE I mean
19:44:19jhMikeSNico_P: looks pretty good so far. I'll have to give a closer look later since I've got some stuff to do around here.
19:54:32ptw419anyone know if there is a way to force a section to be linked even if it isn't used?
19:54:44ptw419i mean functions within a section
19:55:00ptw419w/ gnu ld
19:58:01The-CompilerIt won't work with −−write-partition...
19:58:24Nico_Ppondlife: bufgetdata can return up to the size of the main buffer
19:58:52Nico_Pif the file starts at the beginning of the buffer and fills it
20:02:50pondlifeOK, so the amount of available data.... makes sense.
20:03:07pondlifeI was originally thinking of it as a posix-read-style call.
20:07:27Nico_Ppondlife: what would it return then ?
20:07:37XavierGr bluebrother: I will be around for 5 minutes, I asked you how you managed to get the rtc chips.
20:09:09 Join ender [0] (
20:09:30 Quit ender` (Read error: 104 (Connection reset by peer))
20:09:32 Quit Thundercloud (Remote closed the connection)
20:11:08 Quit lazka ("I'm off now")
20:11:28pondlifeNico_P: The return value would be limited by the length of data requested.. not relevant here though.
20:11:29 Join Thundercloud [0] (
20:11:52Nico_Pok so it's good as it is ?
20:13:25 Join Davide-NYC [0] (
20:15:04Nico_Pafter a few changes it seems to be working a bit... it's "playing back" one of my files
20:15:48Davide-NYCHello again. I just wanted to report here that the test_codec plugin wreaks havoc on my H1x0 and my GBFX. At least when using the "Spreed Test Folder" option.
20:18:11Davide-NYCThe clock gets reset to −−:−− and I am unable to playback anything nor load plugins.
20:21:03 Part Arathis ("Bye, bye")
20:23:26 Join perplexity [0] (
20:24:10 Join kaaloo [0] (
20:24:21 Part kaaloo
20:25:23Nico_Pwoohoo ! it works !
20:25:30Nico_P(on the sim for now)
20:26:04 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
20:27:26Davide-NYCclearing all settings seems to have resolved the problem. False alarm #2.
20:28:23 Join BHSPitLappy [0] (
20:31:32 Quit Covert28 ("Leaving")
20:33:17 Nick hannesd_ is now known as hannesd (
20:35:58gregjwhat's up folks
20:37:56Davide-NYCstill having problems.
20:38:28Davide-NYCtest_codec has problems when using the "Speed Test Folders" option. Where do Ireport my findings?
20:40:08Davide-NYCit's almost like it runs out of memory (or something) at a certain point.
20:41:19 Join BHSPitLappy_ [0] (
20:41:47 Quit BHSPitLappy (Nick collision from services.)
20:41:49 Nick BHSPitLappy_ is now known as BHSPitLappy (
20:42:10 Quit amiconn (" HydraIRC -> <- Would you like to know more?")
20:43:18 Join Biagi [0] (
20:45:37 Join tumu [0] (
20:46:44tumuany update on sansa's bootloader lcd corruption (FS6896) in re: of latest lcd patches?
20:49:41 Part Biagi
20:50:38 Quit Xerion (Read error: 104 (Connection reset by peer))
20:51:51 Join andy__ [0] (
20:53:56andy__Do RockBox support IPod video playback either on the ipod's screen or via TV-Out?
20:53:57 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
20:55:24 Join ]RowaN[ [0] (
20:56:35 Join R2-D2-Master [0] (
20:56:47R2-D2-Masteri need a fast answer
20:57:03 Join lazka [0] (
20:57:10bluebrotherfast answer
20:57:21R2-D2-Masteris 55€ a high price for a used MK3006GAL 30GB IDE 1,8" HDD
20:57:31R2-D2-Masternot ZIF
20:57:43 Join EbErT [0] (
20:59:04bluebrothersearch for mk3006gal on ebay. Look at the second match (ebay shops)
20:59:52bluebrothershould give you at least a feeling ...
21:01:22***Saving seen data "./dancer.seen"
21:01:55 Part maffe
21:03:06R2-D2-Masterthe 1 match is the one i want to buy, the second is a shop where it costs 119 €
21:03:23bluebrotherthat's exactly what I saw :)
21:03:27Soap_55 seems low. used/refurbushed
21:03:29R2-D2-Masteri'm not shure why nobody wants the hdd,
21:03:44bluebrotherno idea
21:03:51R2-D2-Masterbut when it's really not usable
21:04:08 Quit andy__ ("ChatZilla [Firefox]")
21:04:40Soap_(that was supposed to be a question)
21:04:46R2-D2-Masterthan its illegal i think...
21:05:05Soap_what is?
21:05:06R2-D2-Mastereven if it's a private auction (is it the same in englisch?)
21:05:10R2-D2-Masterwhen i buy it
21:05:21R2-D2-Masteran its damaged
21:05:30bluebrotherwhy should it be illegal selling a hard disk?
21:05:33R2-D2-Masterand it wasnt noticed
21:05:40R2-D2-Masterno if its damaged
21:06:08EbErTebay is about trust, ask them if they'll stick to a limited guarantee
21:06:09R2-D2-Masterand the seller didnt write it in the description
21:06:13 Quit entheh (Remote closed the connection)
21:06:35bluebrotherwell, if he doesn't sell it as broken one should assume it's not.
21:06:49pondlifeNico_P: So it works on the sim, what about on target? Or is that still having memmove problem?
21:06:53Soap_I can't read the description :(
21:06:58 Join entheh [0] (
21:07:06Soap_55 Euro is cheap as heck. I bet it is used.
21:07:10Nico_Ppondlife: I think it should work on target... still not tested though
21:07:20pondlifeOK, just wondered.
21:07:30bluebrotherit is described as used.
21:07:46R2-D2-Masterbut not as damaged
21:07:58R2-D2-Masterwhat a decision
21:08:02pondlifeDavide-NYC: Did you find a command-line WMA encoder by the way? This might help:
21:08:04Soap_there is a lot of "professed ignorance" on ebay.
21:09:22Davide-NYCpondlife: thanks. It's not as 'clean' as I'd like. I winder if a single exe file exists.
21:10:53 Join maffe [0] (
21:14:37 Quit Thundercloud (Remote closed the connection)
21:16:06 Join TTSbox [0] (
21:22:22 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
21:22:33saratogathe windows media encoder package comes with a command line encoder
21:23:39saratogathough it doesn't expose anything that not available to the gui
21:25:06 Quit ompaul ("init 6 testing")
21:25:28saratogai'm still trying to come up with a good way to analyze memory accesses in the wma decoder so i can optimize whats in IRAM
21:25:45saratogaanyone tried Valgrind?
21:27:11Davide-NYCsaratoga: is ther an easy way for me to redistribute the encoder without necessitating an installer package and windows validation etc etc
21:27:53Davide-NYCsomething similar in elegance to lame.exe?
21:30:34saratogaDavide-NYC: can you just link to the WME installer?
21:31:29saratogai have no idea if the encoder will work without it installed
21:31:30 Join tiwest [0] (i=5693913f@gateway/web/cgi-irc/
21:32:57pondlifesaratoga (or anyone): What's the best WMA decoding performance seen on Coldfire?
21:33:19saratogapondlife: someone in the logs reported ~120% realtime
21:33:29saratogano idea if theres better, i don't have a coldfire player to test
21:34:28Davide-NYCsaratoga: OK I'm gonna give up on getting a local folder install working
21:35:28Davide-NYCso if I'm going to have users install WME9 what parameters should I use for encoding?
21:35:28 Quit R2-D2-Master ("ChatZilla [Firefox]")
21:35:28Davide-NYC*I meant "local folder non-install"
21:36:25saratogaDavide-NYC: see if this works without it:
21:36:46bluebrotherpondlife, saratoga: I just got 122% on a random test file (192kbps)
21:36:53Davide-NYCis that a modified version?
21:37:04 Quit tiwest ("CGI:IRC (Ping timeout)")
21:37:24saratogaDavide-NYC: no just the stock one
21:37:33saratogaits a wrapper for the DS encoder, so it may work anyway
21:37:44Davide-NYCit no worky (already tried it.
21:37:52merbanansaratoga: what kind of files do you need ?
21:38:12pondlifebluebrother: Is that enought to play without gapping?
21:38:46bluebrotherit worked for me. Haven't checked how that works if it's buffering
21:38:49Davide-NYCsaratoga: I mean it works but the user would need to have installed the package. Which is no biggie, I was just trying to avoid it.
21:39:18bluebrotherbut the audio thread screen showed a boost ratio < 100%
21:43:33 Join hannesd_ [0] (
21:46:00Nico_Ppondlife: I think I should add some screen output before testing on target
21:46:10Nico_Potherwise I won't know what's happening
21:46:37 Join saratoga2 [0] (i=9803c6dd@gateway/web/cgi-irc/
21:47:00Davide-NYCquick question about AIFF (I know, who cares) do we want signed or unsigned data?
21:47:05 Join theBishop [0] (
21:47:21theBishopcan anyone help me with restoring the partitions on my 4G ipod?
21:47:50saratoga2another question: how do i get teh asm for a function out of arm-elf-gcc?
21:48:00saratoga2i can do it for regular gcc, but not the arm cross compiler
21:48:05saratoga2anyone ever figured this out?
21:48:42bluebrotherwell, shouldn't arm-elf-gcc work like the host gcc?
21:49:12 Join PaulJam_ [0] (
21:49:20 Quit saratoga (Read error: 104 (Connection reset by peer))
21:49:26saratoga2bluebrother: yes, but it cannot compile test programs like regular gcc can, only rockbox (or at least things passed through rockbox's makefiles)
21:50:07saratoga2so if I just want to see what the asm of a function looks like, I have no way to tell gcc to just give me the arm asm for that function (though I can easily get the x86)
21:51:07 Quit entheh (Remote closed the connection)
21:51:27 Join entheh [0] (
21:51:28bluebrotheryou could try finding the function by disassembling with objdump
21:51:39n1ssaratoga2: you can add the -S to the compiler switches in the makefile, that will make it output assembler code
21:52:07saratoga2bluebrother: how would I locate the function in the entire codec source?
21:52:33bluebrothermaybe -save-temps helps also?
21:52:54bluebrotherobjdump displays the function names when running on rockbox.elf
21:53:08bluebrotherat least I understand it doing so ;-)
21:53:49theBishopis there a place i could get a stock firmware for 4G
21:53:59n1stheBishop: itunes?
21:54:27bluebrotherobjdump also has an option to intermix source with asm. Haven't figured out how this works if the sources are in subfolders though
21:55:13saratoga2bluebrother: t hanks, that option got it
21:55:27bluebrotherhmm, -I seems to be useful ...
21:55:49 Quit The-Compiler (Read error: 104 (Connection reset by peer))
21:57:01theBishopn1s, i don't run an operating system supported by iTunes
21:57:41bluebrothertheBishop: check the IpodManualRestore wiki page (the page might be named slightly different)
21:57:48 Quit hannesd (Read error: 110 (Connection timed out))
22:00:41theBishopbluebrother, link?
22:00:47 Join jhulst [0] (
22:01:15bluebrotherwell, the Rockbox wiki has a search function ...
22:01:50Nico_Ppondlife: finally I decided to test without adding screen output... I got panic: stkov
22:03:31 Quit PaulJam (Read error: 110 (Connection timed out))
22:09:08 Join Thundercloud [0] (
22:12:03 Nick PaulJam_ is now known as PaulJam (
22:13:30 Join Psilonaut [0] (
22:13:39 Quit theBishop (Remote closed the connection)
22:15:57Nico_Ppondlife: it works on target ! :D
22:16:10 Join dj``bergamote [0] (
22:16:13Nico_PI had to increase the stack size quite a lot but it ended up working
22:16:22n1sNico_P: congratz
22:22:35 Join belze_ [0] (
22:22:59belze_wma is running faster than realtime on my iriver h100 after the last submit :)
22:23:02belze_thx alot guys
22:24:42 Quit TTSbox ("Leaving")
22:25:02 Join stripwax [0] (
22:26:21Davide-NYCguys I need help with a batch file
22:27:04stripwaxgo for it
22:27:16 Join pixelma [0] (i=pixelma@
22:28:06Davide-NYCthe wma encoding lines (line 90) don't work and I don;t know why
22:28:34hcsooh, the pp5020 scaling freeze is fixed?
22:28:38bluebrotherdo they if you run the lines from the command prompt?
22:28:54Davide-NYCyes if I substitute %file% for the filename
22:28:54scorchehcs: since a bit ago thanks to amiconn
22:29:55bluebrotheri.e. %file% becomes sample.wav?
22:30:27Davide-NYCOoops! move the WMCmd.vbs up a level and get rid of the folder "codecs/" reference
22:30:40Davide-NYCcrap let me fix this
22:32:05stripwaxDavide-NYC is this related to this?
22:32:23Davide-NYCyes, but I think me fixed it (maybe)
22:33:05 Nick hannesd_ is now known as hannesd (
22:33:13 Join linuxstb [0] (i=5343d4aa@rockbox/developer/linuxstb) who is going to be willing to patrol the new themes repository beat?...requirements for the job: FTP client, willing to get automated mails notifying you that a theme was submitted and (hopefully) having it be done already, not likely to be hit by a bus any time soon
22:33:19 Quit dj`bergamote (Read error: 110 (Connection timed out))
22:33:57Davide-NYCsaratoga2: what bitrate above 320 does the wma encoder like?
22:36:33 Join Siltaar [0] (
22:36:39 Join amiconn [0] (
22:39:24saratoga2Davide-NYC: I have no idea
22:39:31amiconndan_a: around?
22:39:37saratoga2i don't know what the valid bitrates are for mwa
22:40:08Davide-NYCme neither and I'm striking out just trying different numbers.
22:40:17stripwaxDavide-NYC : - did your m4a tracks really decode in 0.02 s?
22:40:27saratoga2i've been testing with 192k, it only seems marginally slower then 128k
22:40:55amiconnsaratoga2: wma is now usable on coldfire, but it struggles a lot
22:41:11amiconnOn arm it's now completely on par with wma
22:41:16saratoga2amiconn: how big of a difference did those changes make?
22:41:22Davide-NYCI've just finished the batchfile and am updating the wiki page. Hopefully folks will report back with numbers. The test_codec plugin is borked right now
22:41:24saratoga2on ARM they did almost nothing
22:41:48amiconncoldfire: 104%->122% realtime, ipod: 170%->178% realtime
22:42:03amiconnJust for comparison: mp3 is >500% realtime on coldfire...
22:42:08saratoga2thats going to make doing IRAM changes fun then
22:42:17saratoga2since ARM barely notices an improvement
22:42:28saratoga2i almost didn't commit that change because I thought it didn't do anything
22:42:29amiconnThat's not due to arm being arm
22:42:37saratoga2err PP or Sansa
22:42:45saratoga2whatever determines the SDRAM access time
22:42:51amiconnThe pp chips have both instruction & data cache, so that iram doesn't really matter much
22:42:52 Join foleyjd [0] (
22:43:07amiconn(except on PP5002 where it *will* make a difference)
22:43:07saratoga2they have single combined cache i thought?
22:43:29amiconnThe point is: it's for instruction and data
22:43:37saratoga2that reminds me, did anyone ever measure what the cacheline sizes and associativity are?
22:43:48amiconnColdfire only features an instruction cache, but no data cache
22:45:15amiconnThat's why using iram for data on cf is so important...
22:45:50saratoga2i've made a simple (and probably inaccurate) profile of the non-stack memory usage of the codec
22:46:02saratoga2should i just move the common stuff to IRAM and let you test it?
22:47:31linuxstbsaratoga2: My thoughts are that we should try and reduce the memory usage as much as possiblew for the WMA codec - e.g. merge buffers if possible, decode things in-place rather than from one buffer to another, etc. I also think we should try to change it to decode a block at a time, rather than a superframe (which will also reduce RAM requirements)
22:47:56pondlifeNico_P: congratulations, so the memmmove issue has gone away?
22:48:19*pondlife has to go, but will keep an eye on SVNs and the IRC log
22:48:24 Quit pondlife ("disconnected has pondlife")
22:48:36amiconnwma does use an mdct, correct?
22:49:05amiconnMaybe it's similar enough to the one used for mp3?
22:49:07saratoga2amiconn: yes, 5 different ones
22:49:14*amiconn would like to see an emac version
22:49:56saratoga2its the only mdct codec i've seent hat doesn't just do short and long blocks, but rather steps up in powers of two
22:50:21Nico_Ppondlife, if you read the log: The memmove issue is really strange, it only appears in the linked list test function. maybe it causes problems elsewhere but I have yet to see them.
22:50:33*Nico_P takes off
22:50:40amiconnwhat's the issue?
22:51:13saratoga2linuxstb: regarding memory usage, i've been looking at shrinking things that could go into IRAM
22:51:13Nico_Pamiconn: I have some data that appears to be moved incorrectly
22:51:36amiconnI wrote the SH and coldfire versions, the arm version was imported from somewhere afaik
22:51:58Nico_Pamiconn: if you want to look at it, it's in testplugin.c:test_ll, the part I ifdeffed out
22:51:58]RowaN[am i the only one who cant find flip mode in the menus of the sansa port? i saw that it was committed today so was just wondering
22:51:59saratoga2thinking that through more completely, i'm guessing you want to decode one bloack at a time so that you can put the output buffer into IRAM?
22:52:00amiconnmemcpy, memmove and memset were fully tested for correctness
22:52:14Nico_Pamiconn: the code is in Jd's svn: svn://
22:52:40Nico_Pamiconn: I doubt memmove is the culprit, but I can't find the problem
22:53:06 Quit davina (Remote closed the connection)
22:53:10amiconnI would take a look if I wouldn't have so many other things on my todo list :/
22:53:31linuxstbsaratoga2: I'm suggesting to try to reduce memory usage first, and then think about what can go in IRAM.
22:53:36Nico_Pamiconn: don't worry, it's not critical to my progress :)
22:53:40amiconnAnd I'm still undecided about the register setup in the restructured PP clock setup :\
22:54:46*Nico_P takes off for good
22:54:50 Quit Nico_P (Remote closed the connection)
22:55:03saratoga2linuxstb: is there a benefit to reducing memory other then IRAM use?
22:55:03 Join davina [0] (
22:55:33amiconnDo we want maximum flexibility, and calculate the values for each possible clock frequency at runtime?
22:55:51amiconnOr would it be better to go with a fixed set of register values?
22:56:25amiconnThe latter would mean less code, but also less flexibility in changing clock setup
22:57:50amiconnOn coldfire we're doing the latter (but there we're also limited in our choice of clock frequencies - they must be integer multiples of the xtal clock in order to be able to let timers running through a clock change)
22:58:06 Quit davina (Remote closed the connection)
22:58:18amiconnOn PP we can choose freely
22:59:14stripwaxcould set up a fixed set for each increment of 10mhz?
22:59:21linuxstbsaratoga2: Probably not, but IRAM is a big enough reason.
22:59:29pixelma]RowaN[: I don't have a sansa but it should be in settings > general settings > display > lcd settings > upside down (if it's there were it is on the other targets)
23:00:12 Join davina [0] (
23:00:30 Join jhulst_ [0] (
23:00:36 Quit jhulst (Read error: 104 (Connection reset by peer))
23:01:20 Quit lost|X40 ("Leaving")
23:01:23***Saving seen data "./dancer.seen"
23:02:21 Join JdGordon_ [0] (
23:02:51 Quit jhulst_ (Remote closed the connection)
23:02:56 Part n1s
23:03:35]RowaN[thanks pixelma, but thats where i already looked and its not there
23:05:42 Quit tchan ("WeeChat 0.2.5-cvs")
23:06:35 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
23:06:55]RowaN[ah its there now
23:07:07]RowaN[grr.. i had used a build after it was committed.. i dont understand =/
23:09:31 Nick ender is now known as ender` (
23:10:34*ender` yawns
23:10:39alienbiker99did you try redownloading the latest builD?
23:11:16]RowaN[yes, thats what im saying. i just did that again for the sake of sanity and now its there
23:11:18 Quit foleyjd (Read error: 110 (Connection timed out))
23:13:22hcsamiconn: I'm still getting freezes on ipod 4g color
23:14:44hcsnot doing anything fancy, just mp3 playback
23:16:36 Join Arathis [0] (
23:17:46 Quit lazka (Remote closed the connection)
23:18:17 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
23:20:50 Quit davina (Remote closed the connection)
23:22:55 Join davina [0] (
23:23:45EbErTya, i have the same thing hcs, isn't as stable as i would expect by now
23:23:55EbErTipod 60g color
23:27:51Davide-NYCFYI: wma 320 bitrate decoding at 109.08% realtime on H140
23:35:56 Quit bluebrother ("leaving")
23:36:19 Quit Psilonaut (Read error: 104 (Connection reset by peer))
23:37:46stripwaxDavide-NYC - is test_codecs working now?
23:38:32 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
23:38:33 Quit Rick ("I don't need to be here.")
23:38:38stripwaxwiki says it doesn't work. but if it does, i'll do a pp (ipod 5g) test now
23:39:09Davide-NYCI've been messing with it all afternoon. I think it is very delicate. If you separate the filetupes into discrete folders it shouldn't crash
23:39:41Davide-NYCalso don't do anything to the player while the test is running. I froze my H140 by unplugging the charger mid test.
23:39:41stripwaxone folder per extension?
23:39:49*stripwax goes for it
23:39:50 Quit bdgraue (Read error: 104 (Connection reset by peer))
23:41:00Davide-NYCit's a pain because we'll have to compile the results together at the end. I'm just going to run WMA and AAC since those are the ones that run slow currently.
23:41:51stripwaxDavide-NYC - i'll try to run them all (against both latest svn build, and also some of the vorbis + libmad patches in tracker). your encoder batchfile is great btw
23:42:18Davide-NYCwasn't originally mine
23:42:26Davide-NYCI think it was either senab or saratoga
23:42:45amiconnhcs: Somehow I doubt that this still has to do with clock setup... and I can't test as it's 100% stable for me now
23:43:03amiconn(I also have to give back the pp5020 target I used for testing)
23:43:17amiconndan_a: ping......
23:43:58pixelmamaybe it's an arm-elf-gcc version problem again? Just a guess as I don't know much about these things...
23:44:54 Join Rick [0] (
23:46:41 Join Frode_ [0] (n=Frode@
23:47:21Frode_it sounds like the latest build got some hardware noise for the sansa build
23:47:34Frode_did not have any before the update
23:48:19 Join dj`bergamote [0] (
23:50:38Frode_the noise seems to be only when switching tracks
23:50:47saratoga2Frode_: thats always been there
23:50:56Frode_saratoga2, really? hehe
23:51:52Frode_been using rockbox for many months, and i was quite sure it got removed, but i might be wrong
23:52:09 Join Xerion [0] (
23:52:32 Join DerPapst [0] (
23:53:22saratoga2damn it that last SVN commit entered before I'd typed in all my comment
23:53:28saratoga2can someone edit it for me?
23:53:48saratoga2i want to mention Benjamin Larsson for sending me some code
23:57:51stripwaxDavide-NYC: - hm, do I need to do something to make the test_codec.rock show up in "Open With"? Looks like it doesn't by default, at least following those wiki instructions. It's in Plugins, but not Open With, unless I've built incorrectly
23:58:38Davide-NYChmm, all I did was add the test_codec.c line to apps/plugins/SOURCES and recompiled
23:58:49Frode_is there any guides/howtos for setting up visual studio to build rockbox?
23:58:52stripwaxoh. it needs to go to viewers.config ?

Previous day | Next day