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 2007-10-29

00:00:49jhMikeSamiconn: what would be the default font? something chosen per target?
00:01:07markunlinuxstb: check out the end of the file, there you can clearly see it
00:01:12amiconnIf you're talking about the default for the UI font, then yes
00:01:23markunbut you can also see repeating patterns in other places
00:01:42pixelmajhMikeS: hmm, yes. I don't want too many fonts but at least there should be two different available for remote/main display and maybe one in menu/one in wps
00:01:44*jhMikeS think the fixed font is bad on gigabeat...m:robe 500 must be like tiny ants
00:03:17jhMikeSpixelma: I was thinking some stylistic variation could be synthesized on the fly like bold or italic so in effect more could exists without more fonts loaded.
00:03:27*pixelma just wants to point out that not the resolution counts but the pixel size (try looking comparing X5 and M5...)
00:03:29 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:04:06linuxstbmarkun: That's the version that's been unencrypted by the iriver tool?
00:04:25jhMikeSwell, gigabeat is a small screen with small pixels.
00:04:49 Quit bertrik ("bye")
00:05:46jhMikeSe200 and gigabeat are the two I use a larger than normal font on but it's actually more readable with more detail
00:05:59pixelmaI don't think you'll get a nice looking italic effect on the fly (with average Rockbox font sizes) but bold *shrug* or making the padding wider or so... but I'd rate the remote/main screen issue higher
00:06:32amiconnOn-the-fly emboldening should be possible and look okay
00:06:54amiconnThat's something I am planning to play with when done with the low level lcd stuff
00:06:58rasherI still thinks we should select a default user-font sooner rather than later.
00:07:00linuxstbmarkun: I wonder if it's just that decrypt tool not doing things exactly right.
00:07:06LloreanWe could always include some form of algorithmic italics, and leave it up to the users whether it looks too bad for them at that font size or not.
00:07:16markunlinuxstb: yes, it could be that
00:07:45preglowamiconn: that trigger screen really is an abomination, but mainly because how sloppy it is in conforming to the rest of rockbox
00:08:01preglowthe settings system is ok enough
00:08:25rasherThe absolute first thing 99% of users with screens larger than the archoses do is to select a font they can actually read. Why can't we ship with one set?
00:08:31jhMikeSvector fonts ... :) gotta be a reasonable way to do that.
00:08:39 Quit iamben (Read error: 104 (Connection reset by peer))
00:08:49amiconnIt's rather counter-intuitive to operate and it really looks like shit compared to the simple lists used for all other settings
00:08:54amiconnIt's too crowded
00:09:03pixelmarasher: that's the test if they read the manual ;)
00:09:21rasherYou shouldn't have to read the manual to pick up the device and use it...
00:09:44jhMikeSrasher: frankly I agree. a good UI is discoverable.
00:10:24rasherI mean sure, if you run into trouble, you're should read the manual, but that doesn't mean we're excused from making a good effort
00:10:25*amiconn can read the default font on all devices he tried
00:10:29tdtookeI'm almost positive that was a typo since my iPod just froze up when I tried to play something, I'm going to replace DATA_NOT_READY with BUFFER_FULL and see if that works
00:10:33amiconnAlthough it could be better...
00:11:15rasherI think the nimbus font would be a decent choice for a default user-font, since it exists in several different sizes already
00:11:19*jhMikeS hands amiconn a gigabeat running the default font :)
00:11:41markunlinuxstb: look for 5D6983B6 :)
00:11:53pixelmajhMikeS: there won't be huge differences to the Ipod Video which he tried...
00:11:53*preglow votes helvetica :P
00:12:06preglowwe should have someone bitmapize helvetica neue, that would rock :P
00:12:09jhMikeSwhat the res/size on that?
00:12:19linuxstbmarkun: Hmm...
00:12:35markunlinuxstb: every 5th line, besides the first few ends with it
00:12:42amiconnmarkun: That reminds me - any news on that font cache stuff?
00:12:52pixelmajhMikeS: 320x240 and really tiny pixels, a very nice display btw.
00:12:52markunamiconn: no
00:13:18rasherpreglow: Helvetica at smallish sizes looks rather "nervous" to me.
00:13:33rasherpreglow:"> for example.. it looks like pixels are missing
00:13:47rasherLower-case c..
00:13:58 Join Alonea [0] (
00:14:20preglowlower-case c is ugly
00:14:23rasherLooks nicer at 14 and above
00:14:31preglowi'm running the helvetica bundled on my nano
00:14:32preglowlooks great
00:14:41 Join ompaulafk [0] (n=ompaul@gnewsense/friend/ompaul)
00:14:49jhMikeSpixelma: now I recall my sister's video has a quite a larger screen than the Gigabeat but they both have the same res
00:14:54rasherThere's a helvetica in Rockbox? I never knew
00:14:59 Quit ompaul (Read error: 113 (No route to host))
00:15:26pixelmajhMikeS: maybe we should ask linuxstb to compare? ;)
00:15:35*markun likes">
00:15:49linuxstbI just did - they're identical in size...
00:16:03 Nick ompaulafk is now known as ompaul (n=ompaul@gnewsense/friend/ompaul)
00:16:20pixelmaheh, thank you :)
00:16:55jhMikeShmmm...I seemed to remember but then I saw it a few minutes months ago.
00:17:06linuxstbNo, I lied... The video is _very_ slightly larger.
00:17:13*preglow uses helvR12
00:17:46jhMikeSlinuxstb: ok, so I'm not full of it ? whew
00:17:54linuxstbThe gigabeat has a larger black border than the video, but that's the difference.
00:18:07***Saving seen data "./dancer.seen"
00:18:33*jhMikeS doesn't watch the anything could be happening there :)
00:18:59 Join webguest77 [0] (i=51429f93@gateway/web/cgi-irc/
00:19:02*amiconn thinks our device comparison chart should be reworked so that the LCD sizes actually reflect the *active* screen area
00:19:02*preglow wants a customizable status bar
00:19:14*karashata wants consistent font sizes
00:19:15amiconnIt could also include the display's ppi
00:19:26*Llorean agrees with preglow
00:19:37preglowrockbox can look pretty nice these days
00:19:46preglowbut the status bar stands out like a sore thumb
00:19:48*jhMikeS had to use a 10x microscope eyepiece to debug the dither on on gigabeat :)
00:20:08*jhMikeS get's an idea for a "sore thumb" theme
00:20:16*tdtooke thinks typing "/me" sure is popular
00:20:23webguest77hello, how I can svn up to a certain revision please? exemple I want to svn up to rev15347
00:20:30*jhMikeS /me and /me
00:20:32*karashata thinks the default theme is enough of a sore thumb already
00:20:37preglowlord, yes
00:21:25markunwebguest77: type "svn help up"
00:21:27tdtookedoes vmware compile noticeably faster than cygwin, i'm getting old here waiting
00:21:55karashatatdtooke: if I recall what I read correctly, it does
00:21:57jhMikeSumm...not noticeably no...incredibly
00:22:29webguest77markun: I have to go to bet, and I have no time to check, can you please say me the command?
00:22:35tdtookeguess i should switch then, i just went with the first thing i found on the wiki
00:22:54markunwebguest77: why would that be faster?
00:23:08markunsvn up -r 15347
00:23:18 Join Shaid [0] (
00:23:19 Join rep|icant_ [0] (n=rep|
00:23:27webguest77markun: thanks
00:23:32jhMikeStdcooke: me too...then I came here and asked if anything was faster with a day
00:23:51rasherpreglow: yeah, I forgot about that. My father told me it's completely useless to him. As is the vkeyboard.
00:24:07 Join andrewg867 [0] (
00:24:25Lloreantdtooke: VMWare tends to take half the time to compile or less, but can be more frustrating to use for some people.
00:24:40linuxstbtdtooke: I'm running r15351 on my ipod, and playback is fine.
00:25:09*jhMikeS doesn't do fancy things with VMWare. Just rb building/sim testing.
00:25:17tdtookehmm.. I guess I'll have to assume something else is going on with mine then
00:25:43 Quit mirak (Remote closed the connection)
00:26:47 Quit cooz ()
00:26:52jhMikeSpreglow: and I still would like bitmap alpha. the USB screen with a genuine drop shadow would look really sweet.
00:27:23preglowscreens like that would't suffer either, it'd be nice and fast
00:27:29jhMikeSbut that's just a start
00:27:51jhMikeSAlpha can be fast. There's some packed multiply tricks to do it.
00:29:08jhMikeSbesides skipping calcs over the majority of pixels that are either 0% or 100%
00:30:14 Quit rep|icant (Read error: 110 (Connection timed out))
00:32:11 Quit webguest77 ("CGI:IRC")
00:32:46hcsmeh, I'll take a nice high contrast large pixel mono lcd any day
00:33:14lostlogichas anyone had any play stoppages or data aborts on latest SVN?
00:33:40 Quit spiorf (Remote closed the connection)
00:33:45tdtookeman.. I may just get with you guys and push the official build, 'cause I'm finding mine has a set of unique issues I wouldn't wish on a snake!
00:33:57jhMikeSlostlogic: been using my e200 for hours now
00:35:02*karashata hasn't tested the lastest build yet, is going to in a few minutes
00:35:49*tdtooke is reverting the last change he made in his since for some insane reason his crazy patch has finally found a way to brick playback
00:35:53jhMikeSseem like the boost is down by a couple % as well...not sure what would've changed that
00:36:01lostlogicjhMikeS: cool −− I think there might still be one hiding in there but if there is it's hard to tell −− I'm not sure if I was running latest or latest-1 when I got it.
00:37:10lostlogicjhMikeS: also, I think that if we have per-handle mutexes the semantics should be that there is never a handle returned from the list mutex unless its handle mutex is locked
00:37:26 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
00:37:26lostlogicthat would actually resolve what I was worried about in Nico's version of per-handle mutexes
00:38:25jhMikeSactually I thought certain modifications to them should be locked perhaps but they could share locking if a multihandle cache is used
00:39:32lostlogicI'm not clear on the concept of shared locking or handle caching :(
00:39:40 Quit andrewg877 (Read error: 110 (Connection timed out))
00:40:58jhMikeSa simple direct-mapped cache using the proper bits from the id or pointer to have collisions minimal. everything that maps to a particular slot shares the same mutex. not 100% conflict free but should mostly be.
00:41:10 Join ivan`` [0] (n=ivan`
00:41:29lostlogicahh, I gotcha
00:41:45jhMikeSright now, only one handle is cached it looks like
00:41:48preglowamiconn: we don't use a patch for binutils currently, do we?
00:42:06preglowdo you think we should? to handle msac
00:42:06 Join xys [0] (
00:42:16lostlogicjhMikeS: yeah and realistically only 2 or 4 are needed in the cache
00:42:23lostlogicbecause not that many are accessed requently
00:42:42 Quit xys (Connection reset by peer)
00:43:06 Join xys [0] (
00:44:09 Quit tdtooke ("( :: NoNameScript 4.1 :: )")
00:44:20 Quit ompaul ("take it easy out there")
00:44:34jhMikeSwhat we don't want is two frequently used handles in the same slot...if the handle had a slot index an associative scheme could be used. maybe just a bloody hashtable since the secondary addressing scheme only needs doing once. evict the oldest? it just me or did the bass suddenly get a bit of a boost in the recent builds...
00:45:50lostlogicthat's oen approach, the other is to cheat and use the knowledge we have of common usage patterns, which dictates certain things about the frequently used handles
00:46:01markungood night all
00:48:37jhMikeSlostlogic: which types get used the most? audio track?
00:49:40jhMikeSlocking a handle wouldn't need a mutex stay locked anyway. just a reference count or locked flag and release.
00:50:38*jhMikeS borrowed some ideas from the filesystem for the dual core locking scheme
00:52:08lostlogicjhMikeS: yeah, audio data currently
00:52:38lostlogicI think some of what Nico's working on might make the metadata area also need some IO
00:53:54jhMikeSbasically in kernel the heirarchy goes from specific to more general to put processor conflict farther down the chain. Those things can be generalized.
00:56:26 Quit ivan` (Read error: 110 (Connection timed out))
00:57:17lostlogicyeah, as I'm fiddling with some locking semantics ideas here I'm seeing where locking levels would be very helpful.
00:57:43lostlogicread and write don't need to lock against each other, but they both need to be locked out in for a handle to be moved
00:58:13 Quit Rondom ("Ex-Chat")
00:58:20 Join rep|icant [0] (n=rep|
00:58:38 Quit ender` (" Just because I have a short attention span doesn't mean I")
01:00:58jhMikeSlostlogic: The events are a bit in prototype stage. I was considering have them wait for a bitmask flag to be matcheable and could block until all conditions are true (or any). This would allow one object to service multiple locks. It also makes a nice proxy object for doing multiple object waiting (like queue+sem) atomic.
01:02:58lostlogicshares some semantics with IRQ levels then
01:03:54 Quit jumpatrain (Remote closed the connection)
01:03:57 Quit rep|icant_ (Read error: 110 (Connection timed out))
01:04:44 Join jumpatrain [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
01:07:32 Quit Thundercloud_ (Remote closed the connection)
01:10:55jhMikeSFrom what you're saying I gather a handle only needs to be locked when moved and no other time? I wonder about the linkedlist. That's only a next op...hmm.
01:13:37 Part pixelma
01:17:48 Quit petur ("ZzzZzzz")
01:19:41lostlogicjhMikeS: yeah, when a handle is moving it needs to be locked against anything else even looking at it
01:20:36jhMikeSI think find handle should leave the mutex locked given the operations being performed.
01:22:08jhMikeSsnapshot the fd, clear it in the data, then unlock
01:22:33jhMikeSthen the file is exchanged out atomically
01:28:22LloreanIs there any reason dithering should be bad on Gigabeat and not iPod?
01:28:29 Join tdtooke [0] (
01:29:29jhMikeSsay what?
01:29:57 Join BigMac [0] (
01:30:36jhMikeShmmm...last I tried it I had no issue but never tried it with WMA
01:31:19tdtookeWow, after I reverted some changes I made in a patch to deal with a problem that mysteriously appeared before I found that it had resolved itself with a recent revision....neat
01:31:27LloreanjhMikeS: No clue, I haven't done any testing myself
01:32:39jhMikeSHeh...that just makes no sense first thinking about it...except gigabeat doesn't use IRAM.
01:33:06tdtookehere's one for whoever: it is actually possible to brick playback by calling setmargins too zealously in gwps-common.c...
01:34:11jhMikeSif it is there's some serious bad thing going on
01:34:17 Quit gromit` (Read error: 104 (Connection reset by peer))
01:34:32 Quit Alonea ("ChatZilla [Firefox]")
01:36:12 Quit tdtooke ("thanks for you time guys.....seriously")
01:36:38 Join gromit` [0] (
01:37:45 Quit einhirn (Read error: 104 (Connection reset by peer))
01:38:30 Join tdtooke [0] (
01:38:52tdtookeOk, I just registered finally, anybody care to allow me to pm them to test this out?
01:39:56 Join JdGordon [0] (
01:39:56 Join webguest03 [0] (i=4a469ecd@gateway/web/cgi-irc/
01:40:47webguest03i just converted my sansa e200r into a e200 but now the computer won't recognize it. Any help please.
01:41:45 Join midgey [0] (
01:45:21 Quit webguest03 ("CGI:IRC (EOF)")
01:51:56 Quit BigMac ("Leaving")
01:53:05 Join midgey_ [0] (
01:55:23tdtookeis amiconn still here?
01:56:20tdtookeprobably not, if he's british it's probably zero-dark thirty over there
01:56:53 Join webguest68 [0] (i=4a469ecd@gateway/web/cgi-irc/
01:57:22 Join rep|icant_ [0] (n=rep|
01:57:37webguest68i jsut converted my e200r to e200 and i am dual booting with rockbox. For some reason, after i converted it to an e200, my computer won't recognize it. Any help
01:57:53tdtookeok, this one goes out to whoever, has the discussion of viewports in the past entailed a custom list position. Reason I ask is I'm wanting to know if I should abandon that effort or try to make sure it is dependent on nothing and is fit for consideration.
01:58:30Lloreantdtooke: Viewports would allow for a custom list position.
01:58:58Lloreantdtooke: Boiled down to its simplest, a "viewport" is virtually what the screen is physically today: Each viewport can be drawn to as if it were a separate screen. At least if I understand correctly.
02:00:36 Quit xys ("Ex-Chat")
02:00:42LloreanWhich would mean if you set a viewport of 120x320 pixels on a Gigabeat you could then draw the list to it in only half the screen, for example. If I understand correctly
02:00:46tdtookeLlorean: I was thinking that might be what the name implied
02:02:06LloreanIf I get it, all the drawing should work as it does now, except that it's directed to a specific viewport rather than just the screen.
02:03:11 Nick ivan`` is now known as ivan` (n=ivan`
02:03:28tdtookeThat would make for some cool effects, small navigatable playlist in one viewport, equalizer in another, if I'm understanding it right
02:03:49LloreanOne would assume such things would be possible if the controls didn't get too complex.
02:04:38 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
02:04:44LloreanI'd imagine them used for things like overlaying the menu over the bottom half of the WPS during playback, instead of absolute switching between the two screens.
02:05:15LloreanBut I would assume the first step, anyway, is getting the existing code to work with one fullscreen viewport for the main display, and one for the remote.
02:05:25jhMikeSclip regions are needed for that and amiconn's intents weren't to include clipping ones
02:05:56LloreanjhMikeS: I thought the whole point of viewports was that lines would clip and scroll within them.
02:05:57 Quit rep|icant (Read error: 110 (Connection timed out))
02:06:04JdGordonjhMikeS: so we should get it done before he has a chance to complin :)
02:06:07LloreanI'm almost certain that was pretty much their *only* job.
02:06:14jhMikeSjust basic clipping to the rectangle but not complex overlapping. I want full clipping regions...which should be workable.
02:06:29LloreanjhMikeS: Just clipping to the rectangle would work for what I described, wouldn't it?
02:06:51jhMikeSnot if one viewport is partically overlapped by another in front of it
02:07:04 Quit jhulst_ ("Konversation terminated!")
02:07:08tdtookeif this wasn't in C that would be easier
02:07:46LloreanjhMikeS: Aaaah, this is true, but they wouldn't *necessarily* overlap, assuming you destroyed the fullscreen viewport and drew the WPS to a new, smaller one. :)
02:07:48LloreanBut I see what you mean
02:08:09JdGordonLlorean: bugger it.. im not backing down, but we are having a circular argument.. so just commit it and ill urn a patched build
02:08:25jhMikeSthe real problem isn't C, it's 1) retangle band manipulation is rather complicated 2) how to message to draw when one is uncovered?. I have a scheme that can use one rectangle-band region for the whole screen though so memory use would be minimal.
02:09:16JdGordonjhMikeS: if each "screen" has a fucntion to do its drawing, that can be connected to the viewport so telling the screen to draw is no problem
02:09:17jhMikeSLlorean: I think it would be disappointment in the long run if the capability were not put in place.
02:09:20 Quit midgey (Read error: 110 (Connection timed out))
02:09:58jhMikeSJdGordon: but then each widget becomes multithreaded. a queue for the thread controlling would alleviate that however.
02:10:45tdtookeI have a friend that teaches over at LSU, I think I'll ask him his thoughts on overlapping viewports as it were.
02:10:46LloreanJdGordon: I'm still giving it until next weekend. I want to make sure the *only* argument is "we like this way better" and that there's not something else I'm overlooking.
02:11:27LloreanJdGordon: I do plan to give people warning via the list and forums too, probably tomorrow.
02:11:41 Nick fxb is now known as fxb__ (
02:11:58jhMikeStdcooke: it is a bugger. backbuffers for every one are out of the question. it needs to be able to paint based on state at some point for sure.
02:12:24jhMikeS*tdtooke...sorry I keep munging that.
02:12:29 Join bipton [0] (i=476c5230@gateway/web/cgi-irc/
02:12:46LloreanjhMikeS: I think the real problem is static memory allocation for this one.
02:12:52LloreanIn my head, at least.
02:12:52 Quit bipton (Client Quit)
02:12:54jhMikeStdtooke: but ideas, suggestion on efficient implementation would be great
02:13:15 Join bipton [0] (i=476c5230@gateway/web/cgi-irc/
02:13:36tdtookejhMikeS: I don't really have that unfortunately, in the real world I'm a database guy, but my friend actually codes for a living
02:13:37jhMikeSLlorean: Complexity can only be so high for any number of viewports. How many do we want? 16,64, 256?
02:13:45LloreanjhMikeS: 16, I think
02:13:47keanuwebguest68, what'd you do?
02:13:59LloreanjhMikeS: Even our largest screen probably won't be useful after division into even 3 or 4.
02:14:14LloreanUnless you're stacking them and some are completely covered.
02:14:58jhMikeStdtooke: well, you mentioned the LSU friend...I affirming the "yes" to asking. :)
02:15:04biptonfellas where would i ask about a new port?
02:15:22biptonto see if there is any interest from others?
02:16:16jhMikeSLlorean: probably, but division becomes overlap...something like a splash screen or yes/no would just pop up. maybe even a menu. doesn't matter. just like a windowing system.
02:16:47karashatabipton: a good place to start would be the "New Ports" section on the forums, check to see if someone else may have started a thread related to the DAP you're thinking of a port for
02:16:56tdtookeLlorean: I saw you mention a warning in the forums, I guess I'm not following this, what's the warning for?
02:17:26 Join RoC_MasterMind [0] (
02:17:59karashatabipton: also, if you're looking for interest in porting to a DAP, be prepared to help out in any way you can toward it
02:18:08 Quit lee-qid ("aufwiederbyebientotsayonara")
02:18:09***Saving seen data "./dancer.seen"
02:18:20 Quit webguest68 ("CGI:IRC (EOF)")
02:18:28 Join z35 [0] (
02:19:06 Quit BHSPitLappy (Remote closed the connection)
02:19:09Lloreantdtooke: Warning?
02:19:34 Join billenium_ [0] (
02:19:40jhMikeSLlorean: there would be "owned" viewports that only stay attached to a main scrollers etc. They clip strictly to the intersection of the owner's visible part and their own rectangle.
02:19:49tdtookeLlorean: you said something like"I do plan to give people warning via the list and forums too, probably tomorrow."
02:19:50LloreanjhMikeS: Yes, but even then, I don't think we could go that deep. Splash, WPS, "List view" (menus + settings), and maybe, MAYBE Plugin.
02:19:55LloreanThat's about the only four I can think of.
02:20:18Lloreantdtooke: Oh, that I'm changing the keymap on the Sansa.
02:20:37billenium_hello, im running Ubuntu, and i type into my terminal which arm-elf-gcc and nothing happens. do i have to type something different to get arm-elf-gcc so i can compile rockbox?
02:20:52billenium_because in the guide it says type which arm-elf-gcc
02:21:01Lloreanbillenium_: "which arm-elf-gcc" tells you if you have it, not gets it.
02:21:08LloreanRunning the script can get it for you.
02:21:12tdtookeLlorean: ah... silly me. I'm attempting to remap the keys for the iPod on doom myself, I just can't seem to find something that works right without me changing every single thing to accomodate it, and I don't want to do that.
02:21:56jhMikeSLlorean: I'd make it scale anyway so simply changing a #define could increase the number. List view would be a gui widget that's owned by a overlapping viewport.
02:22:09jhMikeSOr the control would be anyway.
02:23:37krazykitbillenium_, if you've already run, you need to edit your $PATH like the guide says
02:24:25billenium_how do i run
02:24:35LloreanjhMikeS: Scaling is good, yeah. Especially when we've got upcoming targets with 640x480 and such
02:24:40jhMikeSI just want to see that if a splash pops up over the WPS, the scrollers and whatever else is happening there don't overwrite it but continue operating behind it with the visible bits still drawing as normal.
02:25:16Lloreanbillenium_: It's just a script in the tools folder
02:25:23biptonkarashata: thanks, i got a microtrack,, it is very nice but their software SEVERELY lacks
02:25:33LloreanjhMikeS: That would be really, really nice. Yes. :)
02:25:54billenium_kinda sorta didnt help at all
02:25:56karashatabipton: you're welcome
02:26:12billenium_how do i run a script thats somewhere in a tools folder?
02:26:45krazykitbillenium_, to be quite honest, if you don't know how to even run a script, i'm not sure compiling rockbox is for you.
02:27:02billenium_ive compiled before
02:27:07jhMikeSLlorean: so one big thing is if two or more viewports operate on the same thread, that one thread must be able to continue handling activity even if the viewport is modal. we have no dispatch system at all.
02:27:14billenium_just not on this OS
02:27:16 Join MethoS-- [0] (
02:27:29krazykitbillenium_, if you don't know how to run scripts, read ubuntu's documentation, as it's really not directly rockbox related
02:28:55jhMikeSheck that even applies without clipping
02:29:15 Nick bagawk_ is now known as bagawk (n=lee@unaffiliated/bagawk)
02:30:17 Quit Nico_P (Remote closed the connection)
02:30:18 Quit bipton ("CGI:IRC (EOF)")
02:30:55billenium_dang this is so hard and confusing...
02:31:08billenium_im not sure if i even have
02:32:15krazykitdo you have the source?
02:32:58krazykitopen a terminal and go there. then cd tools
02:33:15tdtookeI'm not sure I understand what billenium_ is asking, in a linux environment dont' you just type something like "./nameofscript" to run a script?
02:34:23billenium_WARNING: this script is set to install in /usr/local but has no
02:34:23billenium_WARNING: write permission to do so! Please fix and re-run this script
02:35:01tdtooketry: $ chmod ugo+rx nameofscript
02:35:17krazykittdtooke, that isn't the issue
02:35:26krazykitbillenium_, run it with sudo
02:35:42billenium_yes i was thinking about that
02:36:09billenium_might as well get all three, right?
02:37:26 Join tarsius [0] (i=802a5ffd@gateway/web/cgi-irc/
02:37:58tarsiusHello, are there any developers on?
02:38:30tarsiusI've been working on a new port to the Sansa c100 series
02:38:34*karashata would randomly guess that *most* of the people here are devs, though not *all* of them...
02:39:03tarsiusand I just took some photos of my disassembled player that I would like to post in the wiki
02:39:16*tdtooke for instance is a dodgy waster of time
02:40:04tarsiusbut I cannot create a SansaC100Port page in the wiki
02:40:15 Quit PaulJam (".")
02:40:31tarsiusI just created an account, but I need to be added to the TWikiUsers group
02:43:36 Quit tdtooke ("thanks for you time guys.....seriously")
02:43:39 Join tdtooke [0] (
02:45:46billenium_if i go into the bootloader folder and change everything that says .rockbox in the folders to .party will it only recognize the .party folder?
02:46:07billenium_(like inside the .c / .h files)
02:46:31 Quit tdtooke (Client Quit)
02:46:36 Join tdtooke [0] (
02:46:59 Quit tdtooke (Client Quit)
02:48:27 Quit MethoS- (Read error: 110 (Connection timed out))
02:50:33krazykitbillenium_, you'll have to change more than that, probably
02:50:42JdGordonbillenium_: what are you trying to do?
02:52:26 Join hoelz1 [0] (
02:52:40hoelz1hello everyone
02:52:59hoelz1just a quick question: I was wondering why the calendar plugin isn't available for the iPod Nano
02:56:22 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
02:59:23billenium_jdGordon, i dont know. I just thought it would be cool to change it to like Haxor of .haxor ect
03:00:36billenium_now that i think about it... some things might have like run file .rockbox/lang or what ever
03:01:07 Part hoelz1
03:02:30 Quit midgey_ ()
03:03:41 Quit z35 ("Leaving")
03:04:47safetydanbillenium_, there should be a single define that controls that directory name. Look for #define ROCKBOX_DIR or similar.
03:04:51 Join homielowe [0] (
03:04:53safetydanNo guarantee it will actually work though
03:12:26billenium_where might the #define be? xD
03:15:09billenium_oh wow thanks jdgordon
03:15:39billenium_do you think i need to change more?
03:15:55JdGordonthat one and the one in your bootloader and thats it
03:16:00JdGordonwell... should be it
03:16:24 Join rep|icant [0] (n=rep|
03:24:40 Quit rep|icant_ (Read error: 110 (Connection timed out))
03:25:49billenium_sounds good
03:26:24billenium_JdGordon one thing
03:26:41billenium_i only have to change /.rockbox to /.Hackedpod right?
03:27:03billenium_or do i have to change
03:27:17billenium_#define ROCKBOX_DIR to #define HACKEDPOD_DIR
03:27:57 Quit MethoS-- (Read error: 110 (Connection timed out))
03:28:12 Join MethoS-- [0] (
03:28:15billenium_(also that tools script has been running for over an hour, normal?)
03:28:54krazykitbillenium_, the first question, don't change what's in the caps, change what's in the quotes. for the second, it depends. you're compiling everything, i think, and that'll take awhile, expecially on slower machines
03:29:21karashatabillenium_: the define is in place so you only have to change that one entry wherever you'd need to use the directory later on
03:29:59billenium_krazykit, yes im compiling all 3
03:30:14karashataif you change what's being defined, anywhere the ROCKBOX_DIR definition is used would need to be changed to use the HACKEDPOD_DIR definition instead
03:30:17billenium_i didnt think it would take a(n?) hour
03:30:46billenium_wait karashata you are saying the opisite of krazykit i think
03:31:07karashataI could try to put it slightly different terms
03:31:36karashataROCKBOX_DIR is a variable that's being used to define the directory Rockbox is installing to
03:31:53billenium_so wait, doesnt it not really matter?
03:31:59karashatayou only need to change the value of the variable, and not the variable name itself
03:32:07billenium_yes i see now
03:32:24billenium_i could change the variable but it would just be more work
03:32:45karashataso, you can change /.rockbox to /.hackedpod and have it install where you want without changing the variable name
03:33:02billenium_k, i get it now
03:33:28karashatamakes changing where everything's installed easier because you don't need to go change the value wherever ROCKBOX_DIR is being used
03:33:31billenium_then ill try to compile (after changing settings.h in /bootloader and /apps) then see what happens
03:33:41*karashata nods
03:33:43karashatagood luck
03:42:45billenium_viva rockbox
03:44:18billenium_wait... there is no settings.h in /bootloader
03:44:50 Part tarsius
03:45:19 Quit ptw419 (Read error: 110 (Connection timed out))
03:46:27billenium_i have been swindled
03:46:56lostlogicsigh, my attempts to implement stricter locking semantics on buffering have so far resulted in awesome deadlockage
03:49:30SoapLlorean> jhMikeS: Even our largest screen probably won't be useful after division into even 3 or 4.
03:51:07SoapLlorean: If viewports are going to be used to accomplish the multifont / scroll margins / etc text formatting WPS desires - might you not want up to two viewports per line of text on the screen?
03:53:19*karashata thinks WPS building might end up much more difficult with such an implementation, but perhaps is wrong...
03:55:31SoapI don't know why I said "up to two" - but I guess my point is you might very well want multiple viewports per line of text for "way cool super rad" WPS effects.
03:56:56billenium_1 hour and 30 minutes and the /tools script is still going...
03:58:25karashatawell, supposing this does end up implimented I suppose those of us that lime making WPS themes will have lots of fun learning all the new things we can do
03:59:06Soapbillenium_: not knowing the specs of your machine...there is no need to act can take a long time.
03:59:16 Join Davian [0] (n=chatzill@
03:59:55billenium_if i had 300 some mb RAM how long do you suppose it will take?
04:00:14SoapIt took my old computer overnight.
04:00:26Soapw/320 MB of RAM.
04:00:38Soapa watched pot never boils.
04:00:51billenium_how long does it take to compile?
04:00:56billenium_under 30 minutes?
04:01:05Soapbillenium_: what specs are your computer?
04:01:09karashataI've only ever compiled rockbox twice before, and it took upwards of an hour to compile for me in Cygwin
04:01:15billenium_umm 320mb of RAM
04:01:31krazykitkarashata, keep in mind cygwin is especially slow for some reason
04:01:32billenium_30gb HDD =P
04:01:42*karashata knows
04:01:59*billenium_ east karashata
04:02:08 Part Davian
04:02:13karashataI'm gonna look into trying VMWare the next time I feel like compiling a build of my own
04:02:13Soapyes, billenium_, I believe it will take you less than 30 mins to compile in the end.
04:02:26billenium_ill do it before school then
04:02:50SoapI mean, and I think you mean (but I'm not sure) compile rockbox.
04:09:00 Quit MethoS-- (Read error: 110 (Connection timed out))
04:18:12***Saving seen data "./dancer.seen"
04:18:43 Join midgey [0] (
04:24:24 Quit jumpatrain (Remote closed the connection)
04:25:47 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:26:14 Join miepchen^schlaf [0] (
04:26:32 Join jumpatrain [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
04:27:02 Join psycho_maniac [0] (
04:27:33psycho_maniacwhat does "YUV" mean in the lcdframerate wiki page?
04:28:42JdGordoni tihnk its a colour system... or something like that
04:28:42 Quit Rincewind ("Cya")
04:29:03Soapthis is one where wikipedia says it well.
04:29:23Soapyou really need the pretty pictures to envision different colour-space models IMHO.
04:32:45psycho_maniacoh ok i got it now
04:34:22 Quit RoC_MasterMind ("Leaving")
04:48:49 Join Alonea [0] (
04:54:02 Quit Alonea ("ChatZilla [Firefox]")
05:01:58 Quit billytwowilly (Remote closed the connection)
05:03:29 Quit atsea- (Read error: 104 (Connection reset by peer))
05:07:36 Join billytwowilly [0] (
05:11:17LloreanSoap: I suspect even if you're using viewports for WPS positioning, you'll still only end up with 3 or 4 just for the positioning. I think *most* of them could be done with two viewports total for text rendering, plus one for all the images
05:11:31LloreanAt least most of the scroll-margins using WPSes I see simply use them to avoid the album art.
05:11:55LloreanI think a max of as little as 16 would probably still be enough. But of course, I'm conservative in my graphical desires.
05:15:44 Quit bb (Nick collision from services.)
05:15:49 Join bb_ [0] (n=bb@unaffiliated/bb)
05:17:28 Join DarkDaemon [0] (n=Prisoner@
05:17:58DarkDaemoni was wondering if there is a way i can make my ipod 5.5g louder than allowed by its settings
05:18:26krazykitget an external amp
05:18:30scorchenot without massive clipping
05:18:33LloreanYou don't find Rockbox's maximum volume loud enough?
05:18:52psycho_maniacdidnt you ask this before yesterday?
05:19:09DarkDaemoni wasnt here yesterday :P
05:19:25DarkDaemonthe volume is loud enough, i was just wondering if it can go anylouder :P
05:19:34psycho_maniacsomebody else must of asked that then. because i remember this conversation
05:19:41DarkDaemoni have new headphones and i wanna test howloud they can go with out breaking
05:19:55krazykitso you'll know when they break?
05:20:03krazykitsounds like a great plan. you find out, then have to get new ones..
05:20:24psycho_maniacor go deaf
05:20:26DarkDaemonmy old ones broke as soon as i turned it up to full
05:20:32DarkDaemon just got them
05:20:39DarkDaemon30 day return policyu
05:20:50*scorche facepalms
05:20:53psycho_maniacjust hook them up to a reciever or something. those get pretty loud.
05:21:07scorchei before e, except after c...
05:22:01DarkDaemon...and that is?
05:22:37psycho_maniaclike a home theater system
05:22:51DarkDaemonoh true
05:23:02scorcheto stay on topic, without an external amp, or MASSIVE clipping, no it isnt possible
05:25:29DarkDaemonone more question
05:26:00DarkDaemonwhen i hold the middle button and it goes to the playlist menu, is there any way i can move the queue option to the top for easy accessibility?
05:26:17psycho_maniacyou can but its not easy
05:26:20scorcheif you edit the code and recompile, yesw
05:26:37 Quit karashata ("Leaving.")
05:26:38DarkDaemonwhat language is it in?
05:27:00 Join karashata [0] (
05:28:10DarkDaemoni know visual basic, q basic, turning, dark basic and some PIC programming
05:28:14DarkDaemonwould i have trouble with it?
05:28:22scorcheonly you can say for sure
05:28:35DarkDaemongood point
05:28:41DarkDaemonok then i will have a look at the code
05:28:55scorcheit is quite a simple edit
05:29:05DarkDaemonthats what i'm thinking
05:31:44DarkDaemonthanks for the help guys
05:31:48DarkDaemonimma go hit the hay
05:31:52DarkDaemongot school tomorrow -_-
05:32:11 Quit DarkDaemon ("-= TMD-RecruitServer 5.0=-")
05:33:36scorcheugh...too much basic
05:35:02 Join atsea-39 [0] (i=atsea-@gateway/tor/x-ca9a67a726cfc417)
05:50:41 Quit Calcifer (Read error: 110 (Connection timed out))
05:58:48 Quit midgey ()
06:02:16 Join midgey [0] (
06:05:57 Quit jhMikeS (Read error: 110 (Connection timed out))
06:13:33 Join Maxexcloo [0] (
06:15:03 Quit midgey ()
06:18:02 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
06:18:14***Saving seen data "./dancer.seen"
06:18:14 Quit Maxexcloo (Client Quit)
06:19:48 Join ddalton [0] (
06:19:54 Part ddalton
06:37:25 Join rep|icant_ [0] (n=rep|
06:37:51 Nick jhulst_ is now known as jhulst (n=jhulst@unaffiliated/jhulst)
06:45:21 Quit rep|icant (Read error: 110 (Connection timed out))
06:52:40 Join doom27 [0] (i=cc6f7a3d@gateway/web/cgi-irc/
06:52:48doom27anyone online?
06:56:19 Quit doom27 (Client Quit)
07:07:44 Join jumijoze [0] (n=jumijoze@unaffiliated/mike-zed)
07:09:19 Quit karashata ("Leaving.")
07:09:21 Part safetydan
07:14:54 Quit psycho_maniac (Read error: 110 (Connection timed out))
07:17:01 Join donutman25 [0] (n=chatzill@
07:31:03 Quit jhulst ("Konversation terminated!")
07:35:57 Join dongs [0] (
07:35:57 Part toffe82
07:36:00dongshai guyz
07:36:09dongswhats the slowest platform arm7 based that can decode AAC?
07:36:20dongsipud sis 74mhz, how much cpu does aac decode take there
07:39:24dongsanyone know lol.
07:39:28dongsis thi sthing on
07:39:47 Join daurn|iphone [0] (n=daurn@
07:43:59 Join Buschel [0] (
07:44:12scorchedongs: long time no see
07:44:43 Quit daurnimator (Read error: 104 (Connection reset by peer))
07:45:11scorcheif you havent noticed, this channel has gotten much stricter in your would be wise to read the guidelines linked in the topic ;)
07:51:24 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
07:58:42 Quit darkless (Read error: 113 (No route to host))
07:59:23 Quit TMM ("Ex-Chat")
08:00:36dongsscorche: it would be wiser to answer my question so I can just leave
08:01:10dongsand i just looked at the "guidelines" and my question is within them.
08:01:39scorchemy issue was not with the question, but the 7 lines containing it
08:01:47scorchea) we run the ipod at 80 MHz now
08:02:10dongsmy question was whats the slowest system that still supports AAC decode
08:02:12scorcheb) look int he wiki for the test codec plugin, as that will allow you to see for yourself
08:02:22scorcheslowest system as in what?
08:02:34dongssomething rockbox supports that is slow and decodes AAC.
08:03:10scorchewell, all SWcodec devices will play it, and the slowest will be all of the PP devices
08:03:26dongsand how slow is that.
08:03:43scorcheas i just said, 80
08:03:48dongso i c
08:04:19 Join ender` [0] (
08:04:29scorchealso in those guidelines, you will find something about speaking "real english"
08:05:06scorchewhich means that we expect "oh i see" instead ;)
08:05:49 Quit BigBambi (Remote closed the connection)
08:06:08 Join tpickers66 [0] (
08:06:42dongsso how much cpu does aac decode take up.
08:06:56scorche<scorche> b) look int he wiki for the test codec plugin, as that will allow you to see for yourself
08:07:21 Quit Buschel ()
08:07:24dongswouldnt that.. require installing rockbox?
08:07:34scorcheyes, it would
08:07:40dongswell, thats not an option
08:07:58scorchethen you will have to wait around for someone else to do it for you
08:11:31dongsi figured someone in the channel would have this statistic right off the top of their head.
08:11:33 Join hannesd_ [0] (
08:12:20scorchewell, i sure dont, and it doesnt look like anyone else is stepping forward..
08:13:57*JdGordon contemplates buying a c150 to help with the possible port...
08:14:24JdGordonit doesnt have a microSD slot though :(
08:15:36advcomp2019JdGordon, it does have a microsd slot i thought
08:15:39 Quit hannesd (Read error: 145 (Connection timed out))
08:15:40 Nick hannesd_ is now known as hannesd (
08:15:47JdGordonthe c200 does
08:18:18***Saving seen data "./dancer.seen"
08:19:05advcomp2019JdGordon, never mind i might have been thinking of a different player
08:23:31 Quit JdGordon ("Konversation terminated!")
08:24:56 Part dongs ("dongs")
08:25:47 Join JdGordon [0] (
08:27:12 Join Rob2222 [0] (
08:38:17 Join dongs [0] (
08:38:29dongshow big is the aac decode plugin
08:38:53scorchethere is no aac decode plugin
08:39:07scorcheit is a codec
08:39:15dongsoka,y how big is the aac codec.
08:39:43scorchedownload a latest build, unzip, navigate to codecs, profit
08:42:06dongsthat wroks.
08:42:10dongsalso thats pretty huge :(
08:42:21 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
08:44:08 Join Darksaboteur [0] (
08:44:24 Quit Rob222241 (Read error: 110 (Connection timed out))
08:48:19 Join tarsius [0] (i=a807eebb@gateway/web/cgi-irc/
08:51:25 Join linuxstb_ [0] (
08:51:36 Part dongs ("lol")
08:51:44 Quit Darksaboteur ("ChatZilla [Firefox]")
08:54:50 Quit linuxstb_ (Client Quit)
09:09:15 Join petur [0] (n=petur@rockbox/developer/petur)
09:11:09 Join stripwax [0] (
09:14:21linuxstbstripwax: Hi. Do you still have a copy of ihpfirm? Someone mentioned yesterday that the link on the IriverToolsGuide is dead. (I know it's probably obsolete now, but it could be nice to fix the dead link - e.g. attach the source to the wiki).
09:14:55stripwaxlinuxstb - i'll have a look
09:15:57stripwax should do it
09:16:09linuxstbAnd the source?
09:16:15stripwaxno idea where the source code has got to ..
09:17:13stripwaxfound a local copy, i'll upload somewhere but don't have time to update wiki right now
09:17:24linuxstbI'll edit the wiki if you want. - thanks
09:18:52linuxstbThanks - wiki links fixed.
09:19:35 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:21:16 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
09:22:15 Join Zagor_ [0] (
09:22:42 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
09:24:01 Nick Zagor_ is now known as Zagor (
09:25:58 Quit JdGordon (Read error: 104 (Connection reset by peer))
09:27:45linuxstbZagor: I'm curious, did you write the comment at the top of usb_ch9.h that austriancoder committed yesterday? I recognise it from your patch...
09:28:08 Join JdGordon [0] (
09:31:26 Join davina [0] (
09:33:37Zagorlinuxstb: no, the text is from the linux usb/ch9.h header file. only he stripped some bits.
09:33:58BagderI don't like blatant copies from other sources without references
09:34:42Zagorthat's why I kept the lines about linux in that comment
09:35:20Zagorthough I agree it should rather be properly referenced
09:36:38 Join norbusan [0] (
09:36:47 Part norbusan
09:36:52linuxstbZagor: Did you see austriancoder's comment on the USB flyspray task?
09:37:03Zagorno. checking...
09:39:05 Join _FireFly_ [0] (
09:40:49linuxstbLooking at usb_ch9.h in SVN, it seems to be taken from the Linux kernel, with (C) Linux Kernel Developers, and the Rockbox (C) header has been used -
09:40:53 Part _FireFly_ ("Kopete 0.12.3 :")
09:41:27Zagorlinuxstb: yes it is
09:41:56*petur is not surprised, given the fact that he tried to port the Linux stack...
09:44:25linuxstbI'm just thinking of gplv2/v3 issues - by using the Rockbox header, we're changing that file from gplv2 to gpl "any version" (IMO).
09:45:12Zagorlinuxstb: I agree, we should not claim copyright on copied files.
09:46:33BagderI think lots of austriancoder's (c) lines should be checked
09:46:49Bagdersince lots of it is copied from linux and their (c) lines have been removed
09:47:15peturiirc, he was already told not to do that
09:47:26Bagderok, good
09:49:33peturand why bother if his code is going to be removed again anyway....
09:50:58 Join karashata [0] (
09:52:32*Bagder looks at the c100 scan
09:52:48tarsiuswhat do you think?
09:53:04Bagdercould it say ESMT on that second chip?
09:53:11tarsiusit does
09:53:18Bagderah good, then it's a ram module
09:53:29tarsiusyep... i looked up the part number already
09:53:41tarsiusit's posted above the pics
09:54:17Bagderyou mean I need to pay attention as well? ;-P
09:54:31linuxstbtarsius: Could you find the TEA5767 ? I notice you wrote "allegedly"...
09:55:17tarsiuslinuxstb: i'm guessing it's on that little circuit board i didn't remove
09:55:52linuxstbWhat about the main flash module? In my device, that's on a small daughterboard which looks similar to the one on your player.
09:55:58tarsiusthere was a decent-sized chip there, but it was upside down and pressed against the larger PCB, surrounded by the plastic case
09:56:10linuxstbThat sounds like the flash.
09:56:15tarsiusoh okay
09:56:45tarsiuswell i didn't actually see the TEA5767, but it was mentioned on that non-english forum
09:56:47Bagderhere's a pic of a tea5767 btw
09:56:59linuxstbObvious question, but does your C100 have radio? I know the E200 came in versions with and without radio.
09:57:42tarsiusit's highly likely that it's in the C100 because both the m100 (i think the number is 100) and the c100 have FM radios
09:58:38tarsiusso it's probable that Sandisk used the same chip in both... it's looking more and more like the c100 is based on the m100... c=color, m=monochrome?
09:59:26linuxstbSounds likely.
10:00:26tarsiusadditionally, the non-english forum appears to mention it's in the c100, although what they wrote is technically incorrect
10:00:48pixelmathe c100 looks very similar to the c200 from the outside if I find the right pictures
10:01:03linuxstbtarsius: Did you see the commit of my initial work to port to my Telechips device?
10:01:47tarsius( says both the C100 and M200 (not 100... my mistake) have the [ESMT M12S16161A 1M X 16 DRAM]... but the C100 has different ram
10:02:07tarsiusi'm still in the process of getting a suitable linux install
10:02:12linuxstbThose changes should give you a clue about what's needed when it comes to adding the c100 as a target in Rockbox. i.e. everything that's specific to the Logik DAX you will need to do for the c100.
10:03:06tarsiusUbuntu 7.10 (Gutsy) wouldn't boot the livecd, but i did get Knoppix to work... if all else fails, i'll work out of Knoppix
10:04:06 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
10:05:20linuxstbIt shouldn't be that hard to get tcctool working in Windows - once you set things up, Rockbox development is pretty much the same in Windows compared to native Linux (using either Cygwin or vmware).
10:05:44 Join J-23 [0] (
10:06:07peturtarsius: 7.10 is a bit of a mess, I'd try 7.04....
10:06:25*J-23 reboots to Linux
10:06:26 Quit J-23 (Client Quit)
10:06:44tarsiuspetur: i learned that lesson and i have 7.04 74% downloaded ;-)
10:06:47 Join przemhb [0] (
10:07:28*petur did the upgrade and wished he had waited longer to hear reactions first
10:08:36tarsiuspetur: what hardware are you using? i have an HP tx1000 and i've heard HP users are in particularly bad shape with 7.10
10:09:18petura dell D820 laptop - no hardware support issues, just minor annoyances.
10:09:31*petur ends OT talk
10:15:57 Nick fxb__ is now known as fxb (
10:18:21***Saving seen data "./dancer.seen"
10:21:15przemhbHi all
10:26:15 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
10:29:18tarsiustime to install linux. bye all
10:29:22 Part tarsius
10:40:02 Quit MournBlade ("Her lips were wet with venom")
10:40:30przemhbamiconn: Have you read the info about H10's tuner I have send you on Friday? What do you think about it?
10:56:49 Quit rvvs89 (Read error: 104 (Connection reset by peer))
11:02:55 Quit donutman25 ("ChatZilla [Firefox]")
11:02:58 Quit markun (Remote closed the connection)
11:04:27 Join rvvs89 [0] (n=rvvs89@pdpc/supporter/active/rvvs89)
11:05:59 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
11:07:59 Join ApotheoZ [0] (
11:11:47 Join markun [0] (n=markun@rockbox/developer/markun)
11:13:57 Quit TotallyInfected ()
11:14:48 Quit Nico_P (Remote closed the connection)
11:18:39 Join lee-qid [0] (
11:19:55 Quit daurn|iphone (Remote closed the connection)
11:20:59Bagderfor the rockbox port to ipod touch ;-)
11:22:52 Join MajorC [0] (
11:23:23linuxstbIsn't the ipod touch just a PDA?
11:23:54Bagderwell, a dap with pda stuff in it I guess
11:24:13linuxstbNice hack though...
11:24:19Zagorway more suited for ipl than rockbox i'd say
11:24:27MajorCim about to compile my own voice file, is there a good TTS-software that is gpl?
11:24:31Bagderipl is dead
11:24:46ZagorBagder: yeah, well, in principle :)
11:24:56rasherMajorC: espeak, flite, festival
11:24:59linuxstbI wouldn't go that far, but I wouldn't say that's suited either - it's uclinux.
11:25:11linuxstbI'm assuming the touch is capable of running real Linux.
11:25:19markunMajorC: espeak+mbrola is also a good combination
11:25:25JdGordonlinuxstb: bsd...
11:25:38MajorCare they win32 compatible?
11:26:04markunat least festival, mbrola and espeak are
11:26:11linuxstbBeing BSD-based, presumably there's no obligation for Apple to release source?
11:26:11markundon't know about flite
11:26:15Zagorlinuxstb: right I forgot it's uc-based. yeah I mean proper linux would be better than rockbox.
11:26:24billenium_my terminal says |Make your PATH include :/usr/local/sh-elf/bin:/usr/local/m68k-elf/bin:/usr/local/arm-elf/bin
11:26:24billenium_| how do i do so?
11:26:58linuxstbbillenium_: What operating system are you using?
11:27:13daurnimatorbillenium_: PATH=$PATH:/usr/local/sh-elf/bin:/usr/local/m68k-elf/bin:/usr/local/arm-elf/bin
11:27:18daurnimatorexport PATH
11:27:29linuxstbAdd the line "export PATH=$PATH:/usr/local/sh-elf/bin:/usr/local/m68k-elf/bin:/usr/local/arm-elf/bin" to the .bashrc file in your home directory.
11:27:37daurnimatorfor a more permanent solution, add that to your .bash_profile
11:27:55 Join agm3nt [0] (
11:28:10billenium_do i just add it to the bottom of (Both?)
11:28:34daurnimatordo mine or linuxstb's
11:28:42billenium_or both?
11:28:43daurnimatornot both
11:28:46billenium_oh okay
11:29:15billenium_i cant find it
11:29:26linuxstbHow are you looking?
11:29:39linuxstbFiles beginning with a dot are normally hidden by most apps in Unix.
11:29:42daurnimatormake the file yourself then
11:29:56billenium_well i dont know where to look o.0
11:30:04linuxstbType "ls -al" in your home directory.
11:30:06 Quit J3TC- (Read error: 113 (No route to host))
11:30:21daurnimatorhow do you stop a $VAR from expanding when you echo it somewhere?
11:30:48linuxstbPut it in " quotes?
11:30:49daurnimatoreg - echo "$PATH" actually echoes your path bar
11:30:52billenium_oaky i did View show all
11:30:54billenium_and i can see all
11:30:56billenium_and i found it
11:31:17Zagordaurnimator: echo \$PATH
11:31:21linuxstbdaurnimator: Then you want single quotes.
11:31:47daurnimatorthat does it :)P_
11:32:13rasherWhy doesn't add this line to .bashrc? 99.9% of the time it would be correct..
11:32:18 Join TMM [0] (n=hp@
11:32:34linuxstbWhen you run it twice, things would break...
11:32:48rasherit could check if it's already there
11:32:52linuxstbBut it could maybe be smart about it....
11:32:52billenium_i pasted export PATH=$PATH:/usr/local/sh-elf/bin:/usr/local/m68k-elf/bin:/usr/local/arm-elf/bin in the last line of .bashrc
11:32:55billenium_its still not working
11:32:57billenium_when i compile
11:33:00Zagorit could ask
11:33:06rasheralso, it wouldn't really break, would it? Just be confused
11:33:08linuxstbbillenium_: You need to open up a new terminal window.
11:33:27linuxstbbillenium_: When you do that, type "echo $PATH" to see if the changes have taken effect.
11:33:59rasherZagor: or it could write a comment with a magic value in the file as well, and check for that
11:34:27billenium_they have
11:34:34billenium_its all good
11:34:54Zagorrasher: yeah, but I don't like scripts modifying my dotfiles without my permission
11:35:06linuxstbI agree, it should ask.
11:35:38Bagderas long as the question is done first with the others I think that's a good idea
11:36:06linuxstbBut doesn't the script require running as root in the default case?
11:36:27Bagderwith the default paths with a default system, yes
11:36:59rasherin that case I guess it could put it in /etc/whatever
11:37:11 Quit ApotheoZ ()
11:37:28linuxstbAnother option could be to install the compilers in the user's home directory...
11:37:59Bagderthat's probably a good idea in any case
11:38:55linuxstbBagder: Are you planning on writing a build-server kicking interface?
11:40:23rasherPerhaps it could be a file in svn, that when updated triggers the kicking
11:40:36linuxstbI was thinking the same.
11:41:01Bagderbut then I'd have to have a separate process for checking the state of that file
11:41:22rasheror a svn commit-hook!
11:41:40Bagderwell... the svn server is not the build master
11:41:45 Join Thundercloud [0] (
11:42:00Bagdercould still be done, but would take some fun anyway
11:48:09Bagderas a first test
11:48:23BagderI'm making the script exit hard if it detect being run for > 20 mintues
11:48:50Bagderit should be the equivalent of getting killed externally
11:50:25preglowlinuxstb: you talked about allocating seek tables as metadata the other day, yes?
11:57:41 Join ramon8 [0] (
11:58:33 Quit jhMikeS (Nick collision from services.)
11:58:39 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:02:12linuxstbpreglow: Yes, but I don't know enough about Nico_P's work to do it.
12:04:13preglowlinuxstb: nah, just wondered if he commented oin it
12:05:39linuxstbI'm not sure if I mentioned that specifically, I did talk to him about the rest of the metadata struct though.
12:06:09preglowanyway, seems to me we're pretty close to getting rid of the malloc buffer
12:06:18preglowonly two real problem codecs i see are tremor and faad
12:06:24preglowso we should try to go for it
12:06:47 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:11:10 Quit Nico_P (Remote closed the connection)
12:11:42preglowwe should really try to be rid of faad...
12:13:11preglowargh, this helix one seems so nice :/
12:13:37markunif only we could use non GPL (and still open source) codecs with rockbox..
12:14:00preglowdoesn't the helix people claim their license is gpl compatible?
12:14:05preglowhave anyone commented on that at all?
12:14:05Zagormarkun: such as?
12:14:51markunZagor: there is the 3GPP reference AAC decoder
12:14:51linuxstbpreglow: Yes, but IIUC the reverse isn't true, so it's just marketing...
12:14:55preglowsuch as:
12:15:08preglowlinuxstb: right...
12:15:25markunlinuxstb: so rockbox would be shipped under their license?
12:15:26 Join J-23 [0] (
12:16:22J-23Hello! Is TextViewer UTF-8-compatible?
12:16:31Zagorwhat license is it?
12:16:41preglowdoesn't seem like their code has changed much since release
12:16:43linuxstbmarkun: I think it means that if someone combines Helix and GPL code, that person doesn't infringe the Helix license...
12:17:33markunZagor: it doesn't state a specific license, let me look again.
12:18:07markunZagor: it's a fixed point implementation
12:18:17markun(that was a BTW)
12:18:23***Saving seen data "./dancer.seen"
12:18:41markunalso made by (same as FAAD, right?)
12:19:08preglowseems helix would even be able to do sbr in fixed point on ipods
12:19:45markunI'm not sure how fast the 3GPP implementation is
12:20:13linuxstbpreglow: Having said what I've said, I'm not sure exactly what in the RPSL makes it incompatible with the GPL...
12:21:04Zagor"The RPSL is a free software license that is GPL-incompatible for a number of reasons: it requires that derivative works be licensed under the terms of the RPSL, and mandates that any litigation take place in Seattle, Washington." −−
12:21:42preglowwouldn't hurt if ffmpeg got that aac codec of theirs working either
12:21:48*preglow summons merbanan
12:22:13J-23Can I read ISO-8859-2 on Rockbox?
12:22:53linuxstbJ-23: I would expect so - IIUC you need to set the codepage setting to iso-8859-2, and use a font with the required characters.
12:24:01preglowat least the current ffmpeg implementation looks pretty cute
12:24:23linuxstbIs it in the main SVN yet?
12:24:40preglowi wonder why the hell they started all soc work in another branch
12:24:54 Part karashata
12:25:11preglownot much risk of anything happening to it when it's hidden away in another place
12:25:19 Part J-23
12:25:32 Join J-23 [0] (
12:25:50markunJ-23: what's wrong?
12:27:00 Join J3TC- [0] (
12:28:11linuxstbpreglow: Do you think there's any point in further optimising (including maybe restructuring) faad2? I don't understand why the code is fundamentally bad...
12:28:35preglowme neither, i just know that most aac codecs are way smaller than it
12:28:52linuxstbIs there possibly just a lot of dead code there?
12:28:52preglowlinuxstb: seems merbanan has done work on aac.c, he should be able to tell me how far away it is from being usable
12:28:58preglowlinuxstb: maybe, *shrug*
12:29:00 Quit safetydan ("Leaving")
12:29:09linuxstbBut isn't that floating point only still?
12:29:35 Join spiorf [0] (
12:30:11preglowsure, but i'd be more than willing to fix that
12:30:17J-23How I can convert ISO-8859-2 file to UTF-8?
12:30:27 Join rep|icant [0] (n=rep|
12:30:27markunJ-23: with iconv
12:31:03markuncat oldfile.txt | iconv -f iso-8859-2 -t utf-8 > newfile.txt
12:32:27preglowlinuxstb: i've got a feeling it'll be a pretty easy job anyway, for some reason
12:32:53preglowlinuxstb: at least having the usual profile going
12:33:09linuxstbWe would also need to port the codec to Rockbox, which might not be straightforward, depending on how integrated with ffmpeg it is.
12:34:29markunJ-23: was that useful?
12:35:00JdGordondoes anyone know if nico started trying to get cuesheet going with MoB?
12:35:11linuxstbWhat do "SBR" and "PS" mean in AAC? They're defined for our build of faad.
12:35:28markunlinuxstb: PS is parametric stereo
12:35:38pixelmaJ-23: do you want to convert them for vieweing them in rockbox?
12:35:43preglowlinuxstb: both are used in aacplus
12:35:48preglowlinuxstb: only gigabeat manages to decode that
12:36:00preglowJdGordon: he said it was next
12:36:26linuxstbSo is aacplus part of "LC" ?
12:36:33 Quit rep|icant_ (Read error: 113 (No route to host))
12:36:37preglowlc is pure bare-bones aac
12:36:40preglowno sbr, no ps
12:36:43J-23markun: Yes.
12:36:52J-23pixelma: Yes.
12:37:08linuxstbI'm confused by the #defines in libfaad/common.h - LC_ONLY_DECODER is defined, but then so are other things like PS and SBR...
12:37:15markunpreglow: is PS really so expensive?
12:37:22preglowmarkun: i wouldn't think so
12:37:23preglowbut sbr is
12:37:30 Join mrkiko [0] (
12:37:40preglowi've got a really kickass implementation of qmf for speex now, hope that can be reused...
12:37:43*preglow checks out
12:37:56pixelmaJ-23: you don't need to do that, the text viewer supports ISO-8859-2 (according to the menu, can't easily check myself)
12:38:10preglowquadrature mirror filter, a way of splitting a signal in two equally sized bands
12:38:13preglowwith some funny side effects
12:38:17J-23and how I can check codepage of file?
12:38:40markunJ-23: I don't think there is a good way to tell
12:38:53J-23pixelma: Thanks.
12:39:21markunJ-23: otherwise rockbox could do the checking and convert to unicode automatically
12:39:56pixelmathe text viewer has its own codepage setting though and unfortunately this isn't in the manual
12:40:09linuxstbDoes Rockbox check text files for a BOM, regardless of codepage settings?
12:40:11preglowmno, that wasn't the same at all
12:40:13J-23I have tried ISO-8859-2, UTF-8, Win1250...
12:40:33pixelmathe entire text viewer chapter is a outdated :\
12:40:39markunJ-23: do you have a link to the file?
12:40:54J-23but I'm on emulator
12:41:03J-23UI Simulator*
12:41:07linuxstbJ-23: An easy way to test files for character encodings is to open it up in Firefox, and change the character encoding setting in View->Character encoding
12:41:26J-23linuxstb: In Opera too ;)
12:41:36mrkikoI understood many things, but codepages ...
12:42:55J-23File is probably bad
12:43:15J-23or... It's in Win150.
12:43:16markunJ-23: looks like CP1250 to me
12:43:23 Quit MajorC ()
12:44:00J-23but what fonts supports Win1250?
12:44:16*linuxstb gives markun a cookie
12:44:39markunlinuxstb: I actually use w3m
12:45:25markuntextmode internet browser
12:45:37linuxstbAnd it detected the encoding?
12:45:48markunno, just tried a few
12:45:48pixelmaJ-23: see UnicodeFonts in the rockbox wiki (or rasher's page, don't have the URL at hand)
12:46:27J-23but it's Win :|
12:46:48*J-23 converts it to UTF.
12:46:57markunJ-23: it doesn't matter
12:47:07markunthe fonts are not codepage dependant
12:47:20JdGordonany objections to FS #8008 ?
12:49:06 Quit jhMikeS (Nick collision from services.)
12:49:12 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:49:36linuxstbJdGordon: What happens if the track changes whilst you're using that feature?
12:49:47JdGordondunno... ill find out
12:49:56JdGordonI just found it and thought its a great idea
12:50:02markunJ-23: I see CP1250 is not supported, so it does make sense to convert to UTF-8, sorry for the confusion
12:50:22J-23It works! :)
12:50:38J-23hmm, no.
12:50:51preglowparametric stereo was definitely more complex than i expected
12:50:52markunhm, it should be supported, maybe I need to recompile
12:51:02markunJ-23: what doesn't work?
12:51:02pixelmamarkun: I see it in the encoding options in the text viewer...
12:51:08markunpixelma: yes, me too
12:51:10J-23I have only changed font, no downloaded file
12:51:21markunI'll shut up now :)
12:51:30markunJ-23: I just tried it with helvR12
12:52:17J-23It works aldough, TextViewer decodes Win1250
12:52:21JdGordonlinuxstb: well... if you have it open for the current artist and the track changes, the database options stay as the previous artist which I guess is correct
12:52:38J-23but I had bad font
12:52:47J-23now I have set UTF font
12:52:56linuxstbJdGordon: Yes, you would want to to remember the artist at the time you selected it.
12:53:08linuxstbs/to to/it to/
12:53:19markunJ-23: there are not really UTF fonts, but I know what you mean
12:53:29JdGordonso no problems commiting it?
12:53:58linuxstbSounds a useful feature to me, but I can't comment on the implementation.
12:54:46linuxstbSome kind of GUI query editor for the database would be nice...
12:55:26JdGordonsounds like bloat to me :p
12:55:30JdGordonthe vkeyboard isnt good enough?
12:55:46 Join Arathis [0] (
12:55:55 Quit J-23 ()
13:01:17pondlifeJdGordon: If it helps, I'd like to see FS #8008 committed too, just haven't had time to test it/look at it.
13:01:37JdGordoncommiting now.. code looks fine and no problems runninng it
13:05:30JdGordonand onto the problem of the tagnavi.config not being localised... how about we check for langs/tagnavi.<language> on load and shit translated tagnavis in the zip?
13:06:06linuxstbJdGordon: I'm obviously talking about a plugin...
13:06:34JdGordonoh... well that maybe is more possible
13:07:06 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
13:08:30linuxstbAlternatively, it could be done in Rockbox utility...
13:09:05JdGordona tagnavi.config editor on the dap would be more useful
13:09:46linuxstbMaybe... Doing it in Rockbox utility would allow a much nicer UI - and a real keyboard.
13:09:47JdGordonSlasheri: you round?
13:10:22JdGordonnotepad isnt a nice enough UI if you have acces to the computer anyway?
13:10:36linuxstbNot if you're someone like my mother...
13:10:50JdGordonthe defaults arnt good enough for your mum?
13:11:34linuxstbI'm just thinking that a lot of users probably aren't taking advantage of the power of tagnavi, because there's no easy way to edit it. And I don't count a text file as easy for normal users.
13:12:11 Quit jumpatrain (Nick collision from services.)
13:12:48JdGordonyeah ok, I agree
13:12:59JdGordonand BAH... extra yellow :(
13:14:45 Join feng [0] (i=c27f0814@gateway/web/cgi-irc/
13:16:09fengJdGordon: hey. I don't know whether it's possible but it would be great if I could also find the songs in the same folder as the current song. Is it possible to add?
13:16:29JdGordonits gonna be interesting in years to come when the build table is long gone/changed and someone tries to figure out what the heck all the "fix yellow/red" commits are :p
13:16:53JdGordonfeng: I was tinhing the same... I didnt do the patch, but Ill hve a look
13:17:01*pixelma wonders if that is another database feature that doesn't work correctly on hwcodec (as "gather runtime" date IIRC)
13:17:26JdGordonthis one should work fine on hwcodec
13:17:41JdGordongather runtime probably doesnt hook into hwcodec playback properly
13:17:48fengJdGordon: I have some folders with songs by different artists / from different albums. And I like the similar feature in MediaMonkey
13:18:31 Part agm3nt
13:22:05fengJdGordon: hmmm... But wouldn't finding songs in the same folder force a switch to the file browser?
13:22:31*JdGordon has an idea how to add heaps of extra options to this
13:22:35fengJdGordon: then please please implement the feature!
13:23:52linuxstb"heaps of extra options" <- Just what Rockbox needs ;)
13:24:24JdGordon:D just 1 or 2 dozen...
13:24:27JdGordonnothing major
13:24:42JdGordonshould actually simplify that patch a fair bit
13:25:07JdGordonand by simplify I mean give it the ol' Jd touch which could go either way ! :p
13:28:01JdGordonanyone remember how to use offset() ?
13:28:03 Join PaulJam [0] (
13:28:20JdGordonI mean to get a datamembers offset?
13:29:30JdGordonthats the one :)
13:29:57LinusNoff = offsetof(structname, membername);
13:30:34*preglow didn't know of that one
13:30:40 Join roolku [0] (
13:31:06preglowjust basically &struct.members - &struct ?
13:31:50linuxstbIs that C or gcc?
13:31:51preglowwhat's the need for offsetof, then?
13:32:01roolkuJdGordon: thanks for comitting FS #8008 for me. :)
13:32:04linuxstbI guess offsetof is compile-time?
13:32:10preglowgood point
13:32:24preglowoffsetof() is a macro, apparently
13:32:44JdGordonroolku: thats for doing it :) one of the few things i might actually use the DB for
13:32:57roolkuJdGordon: I was going to do it myself, but wanted to remove one restriction (that is still there from the dynamic searches)
13:32:57*JdGordon is currently mangling the patch though :p
13:33:35roolkuJdGordon: currently you can only use one of the dynamic searches in a clause (e.g. the last one), as search string is a dynamic variable
13:33:51JdGordonyou have commit acces? I forgot :p I would have left it for you then
13:33:58roolkugrr...searchstring is a static variable I meant to say
13:34:54roolkuJdGordon: if you want to include this fix in your mangling go ahead. :)
13:35:18JdGordonna, I'll leave that to you... :)
13:35:59 Quit mrkiko (Remote closed the connection)
13:36:54PaulJamJdGordon: does fs#8008 require a rebuild of the database?
13:38:39roolkuJdGordon: are you still going ahead to implement #directory# ?
13:38:55JdGordonim adding all the id3 strings in a very generic way
13:39:04JdGordonwell... thats my aim anyway
13:39:47pixelmaand then go ahead and document it? ;)
13:40:18roolkuJdGordon: okay, so I better wait. There is no id3 string for directory though IIRC
13:40:50JdGordonpixelma: :'( do I haaave to :'(
13:41:11JdGordonroolku: their is the full path which is good.. we just extra handling for it
13:45:44 Quit feng ("CGI:IRC")
13:46:29linuxstbroolku, JdGordon: Wouldn't the code be clearer without the SOURCE_CURRENT_ALBUM and SOURCE_CURRENT_ARTIST #defines? From what I can see, they're only used once.
13:46:59JdGordonits still nicer to have them in a define than hardcoded imho
13:47:44linuxstbNo, because it hides what is going on. IMO, defines are only useful if the same string is used in multiple places.
13:50:09JdGordonok, well, they are being moved more out of the code so I'll replace them with strings then
13:50:20*JdGordon really getting frustarted with the 80char width rule
13:51:53preglowit's there to stay, baby
13:52:17*preglow really loves being able to fit a ton of windows on one screen
13:52:28preglowenough that i even use the 80 char rule privately
13:53:06preglowand i do find it annoying too, but not more annoying than 320 char wide lines that wrap like mad
13:53:36*linuxstb prefers full-screen windows and alt-tab to quickly switch between them
13:54:35preglowi used to do that, but no more
13:54:50JdGordonfullscreen windows with a widesceen lcd is no good
13:55:01Bagderi like to be able to have source code windows side-by-side
13:55:35*JdGordon loves when his code compiles first try!
13:56:11preglowit's especially nice when syncing
13:56:50preglowugh, seems i've broken wma locally
13:58:30JdGordonroolku: still there?
13:59:08 Join norbusan [0] (
13:59:13 Part norbusan
14:00:36*JdGordon has no idea how tagnavi.config works :p
14:02:01 Quit kclaf (Read error: 110 (Connection timed out))
14:02:24 Join MethoS- [0] (
14:03:30 Quit MethoS- (Read error: 104 (Connection reset by peer))
14:12:04 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:14:24 Join webguest24 [0] (i=53ed3c9a@gateway/web/cgi-irc/
14:14:26 Quit webguest24 (Client Quit)
14:15:34 Join Daniel_S [0] (i=57b0db68@gateway/web/cgi-irc/
14:16:10 Join MethoS- [0] (
14:18:27***Saving seen data "./dancer.seen"
14:24:06 Join kclaf [0] (i=kclaf@
14:24:38 Quit Daniel_S ("CGI:IRC")
14:25:28 Join rep|icant_ [0] (n=rep|
14:28:17 Join midgey [0] (
14:29:32 Quit Seed (Nick collision from services.)
14:29:39 Join Seed [0] (
14:30:05 Quit midgey (Client Quit)
14:32:53 Quit rep|icant (Read error: 110 (Connection timed out))
14:40:40 Join parafin|sleep [0] (
14:43:02 Quit parafin|away (Read error: 145 (Connection timed out))
14:43:10PaulJamhi, i noticed that on h300 in mpegplayer there is a long delay (3-4 seconds) between pressing PLAY to pause the video and the video actually pausing. is this normal?
14:44:35 Join Arathis2 [0] (
14:44:48linuxstbI assume that's a new "feature" ?
14:45:42JdGordonroolku: ping?
14:46:18 Quit Arathis (Read error: 110 (Connection timed out))
14:48:20JdGordonany complaints about this diff?
14:49:35PaulJamseems like the delay is caused by a disk spinup. i wonder why the disk has to spin up on pause.
14:51:20 Quit keen- (Read error: 110 (Connection timed out))
14:51:27linuxstbJdGordon: What is "disc" - the disc number (e.g. 1 of 2) ?
14:51:38JdGordonnot 100% sure
14:52:04LinusNPaulJam: to save the position perhaps?
14:52:06linuxstbAlso, does Rockbox support albumartist?
14:52:17JdGordoni tihnk so
14:52:29linuxstbThat may be a useful string to add - maybe instead of disc.
14:52:29JdGordonI just added them to show its doable.. more can be added
14:53:52linuxstbPaulJam, LinusN: Shouldn't it only save the position on exit?
14:54:18LinusNit was just a guess
14:54:50linuxstbAnother reason could be that it's waiting for a buffer to empty. But again, it shouldn't...
14:55:28 Quit hcs ("Leaving.")
14:56:03*linuxstb gives LinusN a cookie - settings_save () is being called...
14:56:16JdGordonwasnt there a commit recently to reduce the time it took the e200 to turn the lcd back on?
14:57:01JdGordonsettings_save() shouldnt cause a spinup.... (unless im forgetting...) it should just register the callback for next time
14:57:11JdGordonunless your using your own settings_sav
14:57:16JdGordonin which case.. ill shutup
14:57:17LinusNJdGordon: this is save_settings()
14:57:25LinusNthe plugin version
14:57:31*pixelma thinks that linuxstb has a lot of cookies to hand out today... :)
14:57:59*linuxstb gives pixelma a cookie too
14:58:10*JdGordon wants a cookie! :'(
14:58:18pixelmathank you :)
14:58:38*linuxstb looks at the empty cookie box and shrugs
14:58:43*pondlife might go to the shops
14:59:03*JdGordon has a temper tantrum and throws random stuff at linuxstb
14:59:21*linuxstb finds another box in the cupboard and gives two cookies to JdGordon
14:59:36JdGordonhehe awesome :D
14:59:45JdGordondoes anyone know the tagnavi.config syntax?
15:00:03PaulJamthere is a wikipage DataBase
15:00:12JdGordonI know...
15:00:13*pondlife hates it when he commits a bad 'un then notices only seconds after the build triggered
15:02:37LinusNno more sucky workarounds for you
15:03:10 Quit Thundercloud (Remote closed the connection)
15:03:15 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
15:03:17pondlifeI blame Nico_P, a little ;p
15:03:22 Join desowin [0] (
15:03:31Nico_Psucky workaround?
15:03:50pondlifeNah, I left some of my debugging code in.
15:05:22Nico_Ppondlife: that was to try to find the cause of the SID segfault, wasn't it?
15:05:31pondlifeYes.. Hopefully a new build will trigger and remove it though!
15:05:47pondlifeI'm still getting a segfault with SIDs though
15:05:59pondlifeDon't know if you saw the Flyspray report?
15:06:06pondlifeLooks like it's in the codec now
15:06:06Nico_PI did
15:06:18Nico_Pyeah I got this segfault earlier
15:06:29Nico_Pdidn't find out why
15:06:47pondlifeThat size=65535 looked suspicious
15:07:10pondlifeBTW, could you maybe make those error #defines all start with ERR_ or something?
15:07:45pondlifeSeeing a 65535 going into arithmetic operations always makes me suspicious.
15:08:15pondlifeSome numbers just shouldn't be involved in arithmetic :)
15:09:15Nico_Pok about the defines... in SID I also managed to get an endless loop at one point
15:12:29 Join kubiix [0] (
15:12:29JdGordonfucking DST... its 1am and im not tired :'(
15:13:37*pondlife still has a load of clocks to put back ;p
15:15:18scorche|wat least yours changed...
15:15:38*scorche|w shakes his fist at congress for having the urge to be "different"
15:15:41 Join kugel [0] (i=kugel@unaffiliated/kugel)
15:18:33*JdGordon gives up.... stupid wiki table sytax... way too annoying for 1am editing!
15:19:08 Join [1]seb [0] (
15:19:13linuxstbIsn't it only 12am really?
15:19:46JdGordonbut its still wiki tables which noone sane can do properly :p
15:20:21 Quit J3TC- (Read error: 110 (Connection timed out))
15:22:05[1]sebdoes rockbox support Micro SD Ultra II cards?
15:22:47linuxstb[1]seb: Are you talking about Rockbox on the E200?
15:23:36JdGordonare they just really fast micro SD cards?
15:24:17[1]sebit thinks so, but twice expensive
15:24:27JdGordonso probably...
15:24:43JdGordonbut why get a fast small card when you can get a 4gb micro for the same price?
15:28:29[1]sebhmm. the fact is there're 2gb microsd for 10, a microsd II 2gb for 20, the cheapest 4 gb comes 40 (microsd hc)
15:29:32[1]sebbut i think 4 gb alltogether is enough for me. i have a 2gb sansa
15:29:55 Join webguest02 [0] (i=52985296@gateway/web/cgi-irc/
15:30:23 Join Rondom [0] (
15:30:30[1]sebso my second question: does the transferrate matter?
15:30:55Zagor[1]seb: not for rockbox
15:31:02JdGordonnope... ive never had a problem with my crappy ultra slow 1gb card
15:31:16JdGordonwriting to it on the computer is spainful though
15:31:56Zagorit would be interesting to see how much difference, if any, a faster card would give
15:33:38[1]sebyou mean writing on it by a cardreader or through the sansa? is there a difference?
15:33:48JdGordonboth are slow
15:34:32Zagori through the sansa. to see if the sansa or the card is the limiting factor
15:34:41Zagori mean*
15:34:47 Join MethoS-- [0] (
15:35:05 Quit JdGordon ("Konversation terminated!")
15:35:07roolkuJdGordon: back
15:35:14 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
15:35:23linuxstbNicely timed.
15:35:23Zagorroolku: that was flawless timing
15:39:08[1]sebok, so i'll think i might pick the fast one. at a near local shop it would get only 1gb for 20 bugs.
15:40:02 Join jgarvey [0] (
15:40:25 Quit webguest02 ("CGI:IRC (Ping timeout)")
15:45:05 Join webguest02 [0] (i=52985296@gateway/web/cgi-irc/
15:45:35 Join J-23 [0] (
15:46:37 Quit kugel ("ChatZilla [Firefox]")
15:48:21 Join Thundercloud [0] (
15:49:25 Part [1]seb
15:50:33 Quit webguest02 ("CGI:IRC (Ping timeout)")
15:51:42 Quit advcomp2019 (Read error: 110 (Connection timed out))
15:52:41 Quit MethoS- (Read error: 101 (Network is unreachable))
15:55:11 Join Redbreva [0] (
16:00:27 Quit Zagor ("Client exiting")
16:00:29 Join cooz [0] (
16:01:46 Part LinusN
16:18:28***Saving seen data "./dancer.seen"
16:26:48 Join toffe82 [0] (
16:26:58 Join mf0102 [0] (n=michi@
16:27:45 Quit J-23 ()
16:29:10 Join J-23 [0] (
16:32:57 Quit markun (Remote closed the connection)
16:36:46 Quit przemhb ("ChatZilla [Firefox]")
16:38:50 Quit J-23 ()
16:42:03 Join markun [0] (n=markun@rockbox/developer/markun)
16:42:46 Part Llorean
16:53:08PaulJamNico_P: i tested the current buld with the uisimulator, and the problem with the runtime data not updating seems to be fixed. But i noticed that in the WPS the autoscore for the first track after starting playback (doesn't matter if resuming or staring playback by selecting a file) doesn't show up. (But i think a similar problem was already present for a long time, but only showed up after skipping a directory)
16:53:19 Nick bb_ is now known as bb (n=bb@unaffiliated/bb)
16:54:15Nico_PPaulJam: cool :) about the autoscore, I think i have an idea... also do you know whether there's an open task about it?
16:55:24PaulJamI'm not sure if there is a fs task
16:57:55PaulJamthat reminds me, i should test if 7318 is still valid.
16:58:54Nico_PPaulJam: do you know how to use gdb?
16:59:13PaulJamsorry, i have no idea
17:01:55Nico_Phmm yeah I think I know what the problem is... I doubt it existed in pre-MoB code though
17:07:26Nico_Pam I the only one finding the DEBUGF output of firmware/id3.c annoying?
17:07:52linuxstbYes, remove it...
17:08:41Nico_PI've long disabled it locally, but using a clean SVN sim reminds me how annoying it is each time
17:10:38linuxstbMaybe they should be changed to LOGF, so that they can be enabled for that file with LOGF_ENABLE (if I understand your last commit...)
17:11:20Nico_Phmm yeah why not...
17:11:31Nico_Pmaybe logf and DEBUGF should be unified anyway
17:13:27linuxstbWe probably need to ask the old-timers why there are two different functions. i.e. if the hardware allows DEBUGF, do you need LOGF as well?
17:14:32Nico_PI agree, even though I hardly see how logf can be preffered over debugf
17:15:23linuxstbI can't think of a reason either. I've never found logf on the target very helpful.
17:17:51*linuxstb is hoping Zagor implements usb serial to enable DEBUGF on portalplayer targets
17:20:59 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
17:23:38amiconnImho both make sense
17:24:02linuxstbBut the question is if we need both at the same time, or could we unify them?
17:24:12amiconnWhile the hardware might allow for using gdb, not all devs will have the necessary equipment, or are willing to open and solder their dap
17:24:57linuxstbI mean it would be a build option to change the behaviour of LOGF to either log to the logf buffer, or to log to the serial port.
17:25:15linuxstbBut there would only be one function in Rockbox.
17:25:36linuxstbSo we don't have a mixture of LOGF/DEBUGF in the code.
17:25:46lostlogicNico_P: I think we should do away with can_add_handle entirely... what do you think of that?
17:28:40Nico_Plostlogic: I was thinking of it
17:30:25lostlogicI mean the cost on bufopen of open and closing the fd is just not that big and the function seems to violate some contract somehow is my thought.
17:34:20 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
17:35:56 Quit scorche (Connection timed out)
17:36:37lostlogicNico_P: I'd also like to change the semantics of move_handle so that it returns a boolean indicating whether the handle is moved and takes a memory_handle** which it either doesn't touch or updaets with the new location of the handle. This makes the caller's work easier, especially if we introduce finer per-handle locking or something of that sort.
17:37:59 Join Mastercactapus_ [0] (
17:38:56 Join Mastercactapus [0] (
17:39:08 Quit Mastercactapus_ (Client Quit)
17:43:06 Part Mastercactapus
17:44:48 Join Buschel [0] (
17:45:38 Join kugel [0] (i=kugel@unaffiliated/kugel)
17:45:57 Quit jhulst ("Konversation terminated!")
17:49:12 Join illissius- [0] (
17:49:49lostlogicNico_P: committed those changes + one other code cleanup, I hope you think they're an improvement as I do.
17:50:09 Join ackbahr [0] (
17:56:19 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
17:56:27 Join Lear [0] (i=chatzill@rockbox/developer/lear)
17:56:50lostlogicgah, screwed it up.
17:57:05 Quit ackbahr (Read error: 104 (Connection reset by peer))
17:57:07LearNico_P: Around?
17:57:19Nico_PLear: kinda
17:57:34LearNico_P: Seen a buffering problem twice now...
17:58:17coozhey guys, is there way to turn off rec button as i constantly keep hitting it by accident
17:58:19 Join webguest93 [0] (i=410848ee@gateway/web/cgi-irc/
17:58:57LearI start a playlist. When there's 20-30 seconds left of the first track, audio stops. UI isn't locked, so I went to the debug screen. PCM was empty, don't remember anything else of importance (sorry).
17:59:21LearSkipping to the next track works, so I skip back to the first track, seek towards the end, and no further playback issues.
17:59:44pondlifeSounds like buffering stalled.
17:59:59pondlifeI had that before the weekend, but it's not happened since.
18:00:30PaulJamNico_P: maybe i was wrong, and ther is still something wrong with updating the runtime data. after having played several tracks now only one got updated. i will observe this and let you know if i can find a pattern.
18:01:09 Quit webguest93 (Client Quit)
18:02:45Learpondlife: does the buffering debug screen show this (don't really know how to read the new one...)?
18:03:08pondlifeWell, all of the bars are empty
18:03:12lostlogicNico_P: my next idea for improving the buffering semantics would be invasive. make any access to the ->next of a memory_handle require &llist_mutex and therfor use a find_next_handle(handle_id) as the common way to iterate the list from outside of the explicit list functions.
18:04:29lostlogicNico_P: (btw, all of the changes I'm introducing ATM are coming about from my _attempt_ to introduce strict locking semantics including per-handle locks that can never be taken up without holding the llist_mutex. I never got that working, but it brought to light the potential concurrency traps in the current system)
18:05:03pondlifeNico_P: Your commit to update the elapsed/offset data smells wrong to me... is it right to update the buffer like that? Plus updating two places each time....
18:05:22lostlogicpondlife: I had the same feeling, but hadn't gotten around to looking at it more closely yet.
18:05:51Learpondlife: IIRC, not all bars were empty though.
18:05:54pondlifeIf curtrack_id3 is having it's stats overwritten, that's the problem.
18:06:07pondlifeLear: Ah, maybe different then
18:06:40pondlifeStraight playback has been working fine for me today.
18:07:09LearWell, I don't have the most recent changes (last 24h or so).
18:08:10jhMikeSlostlogic: I had a look at that and see a way to arrange this with half the mutex usage or less and without locking the list during removal. Only one contractual obligation need be observed and that's if a handle is about to be freed, nothing intends to use it (freeing implies that or else it's like a double-free of memory and is a bug).
18:08:33 Quit illissius` (Read error: 110 (Connection timed out))
18:08:54lostlogicjhMikeS: so your solution would only do strict locking around rm_handle and move_handle?
18:10:25LearBtw, started playback now, and behavior in sim and target is different. In sim, "real" in debug screen is full, on target, it started at 500kB, is now 4MB. This on a Sansa.
18:11:14jhMikeSactually stricter locking than there is now but less
18:11:31lostlogicjhMikeS: sounds good to me
18:12:03lostlogicdoes it use some kind of latch-esque mechanism?
18:12:17 Quit petur ("work->home")
18:12:32jhMikeSfind handle should leave the mutex locked
18:13:00jhMikeSnot sure what you mean by "latch-esque" though
18:13:11lostlogicyeah, that was part of what I did, but I broke stuff.
18:13:29lostlogicI'm still concerned about accessing ->next pointers without llist_mutex held though.
18:14:10jhMikeSif the _first_ operation when removing a handle is to change ->next, it should be fine
18:15:19lostlogichmm *nod*
18:15:26jhMikeSthat's iff the contract above is also observed.
18:15:47lostlogicso you gonna write it up? :-D
18:16:35jhMikeSI could give it a go soon. It seems pretty simple.
18:17:26jhMikeSOne bugger is the callback registration. Should the callbacks be added and removed only on the buffering thread in stead of by the caller?
18:17:48lostlogicthat doesn't seem to be a likely place for a concurrency pitfall though?
18:18:30***Saving seen data "./dancer.seen"
18:19:19jhMikeSdepends...I'm looking at this in terms of utterly simultaneous use to guard against suprise bugs later
18:20:14preglowwhich one should
18:20:28preglowif using threads at all: do it properly
18:20:44 Join Daniel_S [0] (i=57b0db68@gateway/web/cgi-irc/
18:21:07 Quit annulus_ (SendQ exceeded)
18:21:42jhMikeSwe don't need the surpise yield factor creeping in afterall. should have learned the lesson already :)
18:24:03jhMikeSThis mpegplayer's keeping me busy though but it can play the other half of test movies I have now instead of choking.
18:24:12 Join norbusan [0] (
18:24:38 Part norbusan
18:29:29 Join barrywardell [0] (
18:34:35 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
18:40:03 Quit pondlife ("Read error: 110 (Connection slimed out)")
18:45:27 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
18:51:50 Quit idnar (Nick collision from services.)
18:51:53 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
18:55:19 Quit Daniel_S ("CGI:IRC")
18:58:00 Quit Arathis2 ("Bye, bye")
19:00:06 Join bluey [0] (
19:01:52 Join moos [0] (
19:04:25preglowreally, what is the difference between logf and debugf?
19:04:29preglowwhy aren't they one command?
19:04:49lostlogichear hear!
19:05:02Nico_Ppreglow: logf writes to a buffer and debugf is supposed to write to the serial port
19:05:04pondlifeI'd like a combined one, with "channels" i.e. the ability to enable debugging for playback only or whatever
19:05:26pondlifeI thought debugf only output to stdio on sim...
19:05:32pondlifeDidn't know about serial support
19:05:35Nico_Ppondlife: we're not far from it, all that's left to do is turn all the debugf into logf
19:05:36rasherNico_P: they *should* be the same thing, which does the right thing depending on what's available
19:05:45Nico_Prasher: I totally agree
19:05:50roolkulogf can write to serial on h120
19:06:04 Join ilgufo [0] (
19:06:05roolkuit also displays on the remote display
19:06:10pondlifeThat's logf
19:06:10Nico_Pbtw, sorry for disappearing earlier, a friend dropped by
19:06:19Nico_PI'm catching up now
19:06:38roolkupondlife: what is?
19:06:55pondlifelogf writes to the remote, right?
19:07:24roolkulogf is limited to a certain line lenght ( to conicide with display with and array size)
19:07:45roolkudebugf has unlimited (well) line length
19:08:13roolkupondlife: yes, sorry - was a continuation from my previous line
19:10:10preglowroolku: that can be fixed
19:10:44preglowbut i agree completely, they should be the same, logf should behave like debugf if it can
19:10:48preglowif not, then it doesn't
19:11:12preglowpondlife: that enabling should be source level, no channels
19:11:20pixelmaNico_P: what exactly is turned off by your latest commit?
19:11:25preglowpondlife: i don't want any playback.c logging in my binary by default unless i'm actually debugging it
19:11:45pondlifeI'd assume a normal build wouldn't have logging in anyway
19:11:50preglowleaving them in (enabled) in the first place is a pure oversight that ought to be fixed, if you ask me
19:11:51pondlifeOnly a debug build
19:11:58lostlogicNico_P:: is the average bitrate of a file known when bufopen is called? If so it might be reasonable to add bitrate the memory_handle and use the bitrate of the current file in the watermark calculations instead of whatever we default to atm
19:12:00Nico_Ppixelma: the output from the ID3 parser in the sim
19:12:09preglowpondlife: well, sure, but why introduce "channels" for something that should be fixed with ifdefs?
19:12:18Nico_Plostlogic: if the metadata has been loaded then I think it is
19:12:23pondlife#Ifdefs would be fine
19:12:49pondlifeI mean, a more structure way to enable what debug output occurs, at compile time.
19:12:49preglowpondlife: each source file usings logfs should have nice fat #define USE_LOGF at the top
19:13:02Nico_Ppreglow: logf (or even debugf) messages are never included if you don't ask for them to be
19:13:09preglowNico_P: i know
19:13:09pondlifeThat would do. And SVN could have them all commented out.
19:13:15preglowNico_P: but once i want them, they're currently all enabled
19:13:17pixelmaonly the ID3 parser? Btw. in general this info was once useful when I found an aiff that wouldn't want to play in Rockbox...
19:13:21preglowpondlife: indeed
19:13:23Nico_Ppreglow: not anymore
19:13:34preglowNico_P: i don't think your commit dealt with all of them, did it?
19:13:46Nico_Ppixelma: you can reenable the messages easily
19:13:58Nico_Ppreglow: in an earlier commit I disabled all logf messages by default
19:14:15pixelmaI wouldn't have known...
19:14:31preglowNico_P: all of them? i thought we agreed a per-file thing would be best
19:14:47Nico_Ppreglow: they are disabled and per-file enabling is possible
19:14:49preglowif i want to debug tagcache.c, i want to enable a #define in the top of that file
19:15:01Nico_Ppreglow: that's how it's done
19:15:33Nico_Psee r15291 (
19:16:15preglowNico_P: will check it out later
19:17:08*linuxstb would prefer to break the rules and use // to comment out #defines
19:17:46Nico_Pmaybe I should add a note somewhere about how to enable logf output
19:18:01linuxstbpondlife: It just looks tidier to my eyes...
19:18:19pondlifeI think // is tidier all round, but I guess it's what you're used to.
19:18:38pixelmalinuxstb: it's done elsewhere this way (for example for the Ondio backlight mod - got an "ok" there)
19:18:45linuxstbOr maybe they should just be #undef LOGF_ENABLE
19:19:48pondlifeAny reason we have lots of while(1)'s in - rather than while(true) ?
19:20:21 Join przemhb [0] (
19:20:26*pondlife is a type pedant
19:20:35przemhbHi all
19:21:10przemhbbarrywardell: I have some new info about H10's tuner; are you interested?
19:21:36*linuxstb wonders why we have logf (a macro) in lower-case
19:22:23 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:23:09przemhbbarrywardell: tuner/line-in switch is LV44052's B pin, which is connected to GPIOB.2
19:23:25przemhbbarrywardell: LV4052's A pin is -> GND
19:23:52linuxstbpondlife: Lots of C programmers don't use bool...
19:24:01przemhbbarrywardell: BUSMODE of TEA5767 is -> GND so I2C is preselected
19:24:07barrywardellcool. what about enabling power to the tuner?
19:24:59przemhbbarrywardell: VCC is turned on constantly - I did not noticed it to be turned off
19:25:08pondlifelinuxstb: But no objection to it?
19:25:12przemhbbarrywardell: BUSENABLE seems to be...
19:25:29przemhbbarrywardell: left free
19:25:34barrywardelloh, I wonder why I couldn't get communication working in that case. I must try it again
19:25:40przemhbbarrywardell: strange, I now, but
19:26:01przemhbbarrywardell: well - it may work
19:26:48przemhbbarrywardell: tuner's datasheets states that BUSENABLE should be pulled-up to VCC for at least 10us before any data transfer
19:26:52barrywardellprzemhb: can you add that info to this page:
19:27:03przemhbbarrywardell: yes
19:27:27przemhbbarrywardell: I will do this later
19:28:03 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
19:28:04barrywardellfree as in not enabled?
19:28:11przemhbbarrywardell: I was investigating the tuner together with my friend
19:28:19Nico_Plostlogic: reading your commit now, nice
19:28:34przemhbbarrywardell: yes, it seems to be floating
19:28:50przemhbit is not hard wired to VCC
19:29:18barrywardellany idea to pull BUSENABLE to VCC? another gpio bit?
19:29:20przemhbbarrywardell: it is not present on a FPC cable that links daughterboard with mainboard
19:29:40lostlogicNico_P: glad you don't object.
19:30:13przemhbbarrywardell: No. Note that I have connected digital osciloscope and checed if I can catch any positive edge
19:30:35 Join Domonoky [0] (
19:31:01barrywardellprzemhb: so do you think BUSENABLE shouldn't affect the tuner?
19:31:10barrywardellsince it's not connected to the daughterboard anyway?
19:31:17przemhbbarrywardell: full off->power on, power on-> tuner on, tuner on->off <= no edge at all; constantly sth about GND
19:31:56barrywardellpower on -> tuner on : that's in the OF?
19:32:03 Join lazka [0] (
19:32:04przemhbbarrywardell: yes
19:32:07linuxstbpondlife: Objection to what?
19:33:10przemhbbarrywardell: there's a chance that you don't have to give positive edge to the BUSENABLE to enable I2C data transfer
19:34:21przemhbbarrywardell: according to the BUSENABLE simplified schematic in datasheet there is a bipolar PNP transistor with emitter connected trough some current source to VCC and base to BUSENABLE pin
19:34:26 Join Bagder_ [0] (
19:34:45Nico_Ppondlife, lostlogi: about the elapsed/offset commit, I don't really see another solution. I can't really keep a pointer to the write location, can I?
19:35:04przemhbbarrywardell: so it should act the same whether it is connected to VCC or if it is left floating
19:35:21 Join Frazz [0] (
19:35:54przemhbbarrywardell: if only simplified schematic of the input gives enough details
19:36:35przemhbbarrywardell: you have no resistance smaller then 10MOhms to any pin of the mentioned FPC cable
19:36:47amiconnhi przemhb
19:37:06przemhbbarrywardell: it seems that there's no "reset" RC circuit from BUSENABLE to VCC
19:37:15amiconnI found your discoveries in the log, but didn't have time to do anything with them except saving them
19:37:22 Quit SirFunk_ (Remote closed the connection)
19:37:41przemhbbarrywardell: and it seems that there's no C in series to any FPC cable from BUSENABLE
19:37:47przemhbamiconn: Hi
19:37:48*amiconn is currently dissecting the G5.5 ROM
19:38:40przemhbbarrywardell: I thin that we should try enabling H10's tuner basing on the data
19:38:54barrywardelllast time I tried, I don't think i2c communication was working for me.
19:38:54barrywardellso I must be missing something
19:39:18przemhb(sorry for the missing letters - keyb. problem)
19:39:30przemhbbarrywardell: what?
19:39:52przemhbI think we have to assume that we do not need to drive BUSENABLE
19:40:20 Join J3TC- [0] (
19:40:31przemhbin case we would be wrong there's a second option which should leed us toward our aim
19:41:09lostlogicNico_P: I don't understand fully the need to update the mp3info on the buffer at all −− how does the runtime data end up reading from it?
19:41:25lostlogicNico_P: my thought is that it should only be updated once it's become static (ie the track it belogns to is no longer being played)
19:42:13Nico_Plostlogic: the buffer/unbuffer callbacks access the handles in the main buffer
19:42:51Nico_Psee audio_generate_postbuffer_events() and audio_clear_track_entries()
19:42:59 Join Daniel_S [0] (i=57b0db68@gateway/web/cgi-irc/
19:43:12lostlogicNico_P: gotcha, then we should offer playback a callback that can update them when they are no longer playback's current track.
19:43:35przemhbbarrywardell: I have written a "set_gpio" plugin, which I have used to chec which GPIO pin drives LV4052 switch
19:43:50Nico_Plostlogic: yes, maybe
19:43:57barrywardellprzemhb: sorry, I'm on a bad connection here. I'm slow in receiving IRC messages.
19:44:06roolkulostlogic: depends what you mean by update - I think the idea is not to spin up the disk
19:44:10 Quit J3TC- (Client Quit)
19:44:12Nico_PI want to rethink the buffer/unbuffer callbacks too
19:44:26pixelmalostlogic, Nico_P: could that fix the "gather runtime data" not working on hwcodec?
19:44:37lostlogicroolku: it's not on the disk, just in memory.
19:44:40przemhbbarrywardell: as BUSENABLE wasn't present on FPC, I didn't try the plug to find GPIO pin which could drive BUSENABLE
19:44:47Nico_Ppixelma: that's probably broken elsewhere
19:44:53barrywardellprzemhb: I'll dig out my old patch later and try it again. I'll post it to the tracker too.
19:45:12przemhbbarrywardell: what patch?
19:46:15Nico_Plostlogic: can you tell me more about your locking ideas?
19:46:19roolkulostlogic: that's why I wrote "depends on" - I am fairly sure it currently updates to disk (certainly if tagcache not in ram)
19:46:33 Join merbanan [0] (n=banan@
19:46:59Nico_Proolku: IIUC the updates are queued and probably only written on the following disk spinup
19:47:00 Join mirak [0] (
19:47:02lostlogicroolku: the runtime data stuff would still hafta hit disk, but all I'm talking about is how and when the data gets from teh playback side of the fence back to the buffer side
19:47:06przemhbbarrywardell: in case I've missed BUSENABLE driving edge (I do not think I did) there's still an option with my plug or testing pin by pin driving them for 10us L->H->L
19:47:07 Join J3TC- [0] (
19:47:28lostlogicNico_P: I have a no-longer-applies (and broken any way) patch that might explain what I was after... I'll post it.
19:47:49 Quit jhulst ("Konversation terminated!")
19:47:58Nico_Plostlogic: did I miss some recent discussion?
19:48:20Nico_PI saw one between you and jhMikeS invloving caching multiple handles
19:49:04lostlogicNico_P: Mike has an idea that involves some similarity with what I suggest (holding a lock on erturn from find-handle that the caller must then unlock when they are done with the found handle)
19:49:20lostlogicI'm pretty confident that his idea is right and mine is just excessive caution at the expense of sanity
19:49:48 Quit TMM ("Ex-Chat")
19:50:03Nico_Plostlogic: what's the basic of your idea?
19:50:25Nico_Plocking both the linked list mutex and a per-handle mutex?
19:50:26barrywardellprzemhb: I previously tried to enable the H10 radio with rockbox. I have a patch somewhere saved on another computer which I'll have access to later
19:50:53Nico_Poh yes, find_next_handle...
19:50:55 Quit Bagder (Read error: 110 (Connection timed out))
19:51:10lostlogicNico_P: the basic idea is that a handle can never leave the llist_mutex without its own mutex being locked
19:51:10barrywardellprzemhb: that set_gpio plugin could be kinda useful too. do you have a patch?
19:51:53lostlogicthe locking semantics of that are complex to realize though because they require strict adherance to order of lock uptake and strictly well behaved clients to the linked list functions that will always drop their locks as appropriate.
19:52:00przemhbbarrywardell: no I don't have, but I can send it to the tracker
19:52:07lostlogic(which is why that patch is large and unfunctional)
19:52:08 Quit Daniel_S ("CGI:IRC")
19:52:13przemhbbarrywardell: I will prepare the patch
19:52:31przemhbbarrywardell: currently it is for H10 only
19:53:20przemhbbarrywardell: it make use of the knowledge about default GPIO's states
19:53:36Nico_Plostlogic: yeah it seems a bit invasive
19:53:45przemhbbarrywardell: it enables to change value of a selected GPIO
19:54:08przemhband it's direction, but only from I/O to O
19:54:20lostlogicNico_P: it was never intended for commit, more as a "is this possible and what traps does it reveal" thing
19:54:38przemhbbarrywardell: for safety reasons there are masks which prevent setting pin which byy default is an I as an O
19:55:00barrywardellprzemhb: thanks. It might be helpful for all PortalPlayer devices
19:55:05przemhbbarrywardell: as it could lead to O-O conflict and could possibly damage logic
19:55:51przemhbbarrywardell: of course it cannot be commited as it's safety is limited
19:56:04Buschelamiconn: you there?
19:57:32Buschelamiconn: great. i saw you did some further rework of the 5G LCD driver and i could not let my hands off it :o)
19:57:38przemhbbarrywardell: will you be able to send your patch this evening?
19:57:58 Join bluebrother [0] (i=tS0Jksml@rockbox/staff/bluebrother)
19:58:19amiconnBuschel: Imo we should neither remove the bcm wait nor drop frames. That's dirty and leads to graphical artifacts
19:58:33jhMikeSAre displacements going to be used in metadata stucts instead of pointers in the future? Not having to adjust them when they're moved would be nice. (things like adjust_mp3entry)
19:58:34amiconnInstead we should try to get a better understanding of the bcm
19:58:45amiconnIt looks like it is capable of double buffering
19:59:00Buschelamiconn: did play around a bit and got it (some little) faster. interesting is that you can reduce the bcm-commanding by far
19:59:03Nico_PjhMikeS: could be good, yeah
19:59:19Buschelamiconn: lots of the register writes/read seem to not have any effect
19:59:28Nico_Plostlogic: did it reveal traps?
19:59:37barrywardellprzemhb: yes, in a couple of hours.
19:59:53amiconnBuschel: The sequences are obviously very close to what the G5.5 rom does
20:00:02Nico_PjhMikeS: linuxstb also raised the idea of having a separate buffer for the strings
20:00:03Buschelamiconn: about the frame-dropping -> not the best solution, but without any bad effect for video playback. i did never see any artifacts
20:00:46amiconnThe 0xfffa0005 is somekind of magic or command; I found some other in the rom, but don't know yet what they do
20:00:55lostlogicNico_P: my biggest concerns that it indicated are: unlocked access to first_handle, cur_handle and h->next and the other two that I already fixed (the potential escape of a h with move_handle and the can_add_handle questionable safety)
20:00:56preglowit could be a nice option, it's fairly common: vsync on/off
20:01:34lostlogicNico_P: btw, it's not necessary to update buf_widx as often as we currently are afaics −− probably sufficient to update it as part of the update_data_counters call, which I think is safer.
20:01:37przemhbbarrywardell: great
20:01:42amiconn(namely 0xfff5000a and 0xfff70008)
20:01:59przemhbbarrywardell: maybe I'll be able to take a look at it today
20:02:21amiconnAnother interesting area would be the investigation of how to use the on-disk version of the bcm firmware
20:02:34amiconnIt seems to be quite different from the rom version
20:02:36jhMikeSNico_P: now exceptions start creeping in?
20:03:25Buschelamiconn: did you try to set the other magic values to x1F8?
20:03:45amiconnOnly one of them; 0xfff5000a
20:04:16amiconnBut they are obviously having a different purpose than 0xfffa0005
20:04:37amiconnUsing the other one leads to 100% blind operated rockbox...
20:04:45Nico_PjhMikeS: exceptions?
20:05:08amiconnAnd the ROM routines using these constants do not set up the update rectangle etc like with 0xfffa0005
20:05:29 Quit J3TC- (Read error: 110 (Connection timed out))
20:05:31amiconn(but they send them to reg 0x1f8)
20:05:39jhMikeSNico_P: this format gets a separate buffer, this one doesn't, etc. etc. IMHO any data format used on buffer should be moveable by design.
20:05:58lostlogicjhMikeS: I agree, I've always hated the absolute pointers in struct mp3info
20:06:46 Quit moos ("Glory to Rockbox")
20:08:11Nico_Poffsets wouldn't allow the const genres array though
20:08:27preglowpointers, ugh, nast
20:08:44Nico_PjhMikeS: a pointer is needed for those
20:08:53Nico_Pbut I still agree offsets are way better
20:09:05jhMikeSyeah, but those string don't move on the buffer
20:09:14 Nick Buschel is now known as Buschel_off (
20:09:23 Join J3TC- [0] (
20:09:38Nico_PjhMikeS: yes but you can either have a pointer to a const array or a pointer t the inner buffer
20:09:54amiconnBuschel: Ah, now I understand the basic principle of these command words: the second halfword is always the bitwise inverse of the first (some kind of consistency check)?
20:10:13Nico_PjhMikeS: it's easily solved by using a "genre" int though
20:10:17jhMikeSah, right. always flagable somehow
20:10:18amiconnI found some more: 0xffeb0014 and 0xffec0013
20:10:46jhMikeSthat's good too
20:10:48Nico_PjhMikeS: this const array was precisley the cause of a crash. the pointer was adjusted when it shouldn't
20:11:39 Join TotallyInfected [0] (
20:12:05jhMikeSaha. those little oopses always creep in with big reworks.
20:12:33Nico_Pyeah, it was a long-standing bug revealed by MoB :)
20:12:44amiconnBuschel_off: And the 0x31 is some kind of 'go' command'
20:13:02jhMikeSnot much different than leaving out an interrupt enable in 3000 lines of kernel code by mistake
20:13:10PaulJamNico_P: i now did a little bit of testing on my h300 and the gather runtime feature is still broken. In the wps the runtime data is missing for the first track and the tracks that are played after rebufering (if rebuffering happened during track 6, then playcount and autoscore are missing for track 7). and the runtime data doesn't get updated for every track (only for first and last track it seems to get updated reliably)
20:14:39Nico_Phmm I think the callbacks need to be reworked
20:15:34*Nico_P has to go
20:16:32 Quit Nico_P (Remote closed the connection)
20:17:21 Quit desowin ("use linux")
20:17:43 Join darkless [0] (
20:18:32***Saving seen data "./dancer.seen"
20:18:56jhMikeSlostlogic: my first round: find_handle() locks the mutex, tries to find the handle. if not found unlocks it. if returned, mutex stays locked. some other functions adjusted to take a handle pointer instead of an id (which implies locked). this avoids additional searches when the pointer is already available.
20:19:22 Join Fraser [0] (
20:19:51lostlogicjhMikeS: cool, that's kinda the other (and almost assuredly better) approach from what I was doing −− I was passing IDs only most of the time forcingn everyone to re-find and re-lock.
20:20:19 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
20:22:05 Join abstrak [0] (
20:22:05jhMikeSthe rebuffer_handle function is only called in one spot and makes it harder to follow. it's critical to follow because it uses send_message inside the lock. mind if that goes?
20:22:33 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
20:22:54jhMikeSI'm just thinking to move the few lines of code straight into bufseek
20:23:14lostlogicjhMikeS: that's fine with me −− functions should enhance readability, not the othe rway around.
20:23:19 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
20:23:23lostlogic(I had considered doing the same)
20:24:20 Join ShadowCode [0] (n=aa@
20:24:27ShadowCodehi all
20:24:36ShadowCodeI have a question :)
20:24:36 Quit J3TC- (Read error: 104 (Connection reset by peer))
20:25:15 Join J3TC- [0] (
20:25:19preglowShadowCode: then you must ask it
20:25:23ShadowCodei have a zsp400 DSP core based rockchip player
20:25:37ShadowCodeand i want to port Rockbox on it
20:25:49ShadowCodecen you qive me some starting help
20:27:12DomonokyShadowCode: take a look at the NewPort wiki page...
20:27:40ShadowCodeyes but the info ther is not helping me
20:28:02preglowhm, 16 bit dsp
20:28:06ShadowCodei need some one to tell me from where in the code to start
20:28:07preglowlook forward to a hard port
20:28:23Domonokyis there a compiler for this chip ? :-)
20:28:32preglowShadowCode: you need documentation first, then compilers
20:28:34ShadowCodeit is RK1608A 100MHz CPU
20:28:40preglowin other words, what the newport page says
20:28:45preglowyou don't start coding the first thing
20:28:49ShadowCodei have the data sheet
20:29:24preglowi can't seem to find any info on any rk1608a cpu
20:29:31ShadowCodeyes i am not new to rockbox but i need guidense
20:29:43ShadowCodeI dont know about the compiler
20:29:46DomonokyShadowCode: if you have a compiler... the next step would be to find out how to start your own code on it..
20:30:18abstrakAny TCC78X datasheet here ? (want to port D2)
20:30:19ShadowCodethat is the hard part
20:30:31ShadowCodei dont now how to test the code
20:31:03ShadowCodei have a disassembler for the games of the player
20:31:06Domonokyabstrak: i think there are TCC datasheets "somewhere".. :-)
20:31:25ShadowCodemaby the asm instructions cen be usfule ?
20:31:25abstrakDomonoky: hard to find...
20:31:32preglowdon't we have someone porting to D2? TMM?
20:31:42ShadowCodeD2 ?
20:31:47preglowiaudio, no?
20:32:09Domonokythere is someone working on D2..
20:32:17abstrakwho ?
20:32:17ShadowCodewhat is D2 ?
20:32:29preglowdomonoky: would that someone be tMM? i can't remember...
20:32:33preglowShadowCode: an iaudio player
20:32:44Domonokypreglow, yes i think so...
20:32:50 Join lukaswayne9 [0] (
20:32:58preglowabstrak: then you should try getting hold of him
20:33:02ShadowCodestrange thing :)
20:33:33 Join linuxstb_ [0] (
20:33:35 Join donutman25 [0] (n=chatzill@
20:33:49Domonokythe TMM is also the person to aks for the datasheets, but maybe someone else also have them.. :-)
20:33:55ShadowCodeok dose he come on irc channal ?
20:34:02linuxstb_TMM is working on the iaudio 6/7, not the D2.
20:34:05 Quit lukaswayne9 (Client Quit)
20:34:09ShadowCodeI have the Datasheet
20:34:40DomonokyShadowCode: dont mix up.. i was taking to abstrak :-)
20:34:59ShadowCodesry :)
20:34:59abstrakShadowCode: u can send me ?
20:35:07preglowlinuxstb_: hmm, anyone working on d2?
20:35:18ShadowCodew8 a sec
20:35:33 Join Llorean [0] (
20:36:09linuxstb_preglow: Not afaik.
20:36:16preglowmok, then i'm just confused
20:36:41linuxstb_Although tcctool has been successfully tested on the D2, so we can safely upload and run code via usb boot.
20:36:53 Quit Frazz (Connection timed out)
20:36:56 Join Frazz [0] (
20:37:36ShadowCodeok that is something i like
20:37:50ShadowCodeis the tcctool on the site ?
20:38:21linuxstb_It's in the Rockbox SVN - in utils/tcctool/
20:39:20ShadowCodeso this tool cen run code on the device via usb ?
20:39:47 Quit sd (Nick collision from services.)
20:39:51 Join sd__ [0] (n=sd@
20:40:28linuxstb_Telechips CPUs have a usb boot mode. tcctool implements the PC-side of that process, transferring a firmware file to the device, which is then executed.
20:40:46 Quit Fraser (Success)
20:41:11 Quit Lear ("ChatZilla [Firefox]")
20:41:28Domonokyabstrak: please join #datasheets :-)
20:42:28ShadowCodesry Domonky i dident see the name :)
20:43:51ShadowCodeand haw do i get in usb boot mode the only mode i know is upgread and i get there with shorting the cpu :)
20:44:05preglowShadowCode: not all players have any usb boot mode
20:44:11preglowwe're talking iaudio d2 here
20:44:21 Join bertrik [0] (
20:45:10ShadowCodeI am asking about the rockchip rk2608 players
20:46:38preglowwell, i don't think anyone here has experience with it
20:47:07ShadowCodehm this is vary bad
20:47:22ShadowCodethis player have a very good CPU
20:47:32ShadowCodeto bad :(
20:48:34ShadowCodeTeclast C260 is the official name of the player :)
20:48:47ShadowCodehave you seen it ?
20:48:51linuxstb_abstrak: Have you seen the Cowon D2 thread in the Rockbox forums?
20:48:58bertrikShadowCode: what kind of CPU is in it? do you have a link?
20:49:31 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
20:49:51abstraklinuxstb: yep
20:51:07abstraklinuxstb: i read it every ~2 days... not a lot of interested people
20:51:28linuxstb_I think there's a lot of interest, just not anyone who is both interested and skilled.
20:52:08linuxstb_But if you can get past the touchscreen, it looks a nice player.
20:52:18*preglow feels like aquiring a telechips target
20:52:24preglowno touchscreen...
20:53:48preglowlinuxstb_: what tcc does iaudio 6/7 use?
20:54:21abstraklinuxstb_: it's my first rockbox port, i cannot port it alone but with a little help to start...
20:54:43linuxstb_The difference between the 77x CPUs seems to be which combinations of modules are included - the list is (I think) 2MB SDRAM, NOR flash and audio codec.
20:55:10linuxstb_The iaudio 7 has an external SDRAM module, so presumably the 771 doesn't have the builtin SDRAM.
20:55:15ShadowCodelook in the forum
20:55:28ShadowCodeyou cen find pics on google
20:55:31linuxstb_My device has no external SDRAM (773L)
20:55:42 Join Alonea [0] (
20:56:20linuxstb_The 7801 is a different beast - dual ARM cores (arm926 and arm946)
20:57:31linuxstb_Plus DSPs...
20:58:15preglowwhat has that? :>
20:58:33linuxstb_abstrak: A lot of the work going on with other Telechips deviecs will help a D2 port - tcctool is the first example, but a lot of the code should be close to what you need as well.
20:58:43linuxstb_preglow: Cowon D2 and iriver Clix2
20:59:01 Quit ilgufo (Remote closed the connection)
20:59:13 Join xys [0] (
20:59:38 Join ilgufo [0] (
21:00:10ShadowCodelook this its c260
21:00:56abstraklinuxstb_: what's the steps ? the first thing to code/port ?
21:02:07linuxstb_You basically need to write drivers for all the bits of hardware. It makes sense to start with the LCD, as that makes everything easier.
21:02:22*abstrak still want to find datasheet (TCC77X/78X)
21:03:13*abstrak is still on #datasheets
21:04:10linuxstb_To do that, you'll need to disassemble the original firmware, and try and find the LCD driver functions there.
21:04:11 Quit barrywardell ()
21:05:20 Part pondlife ("Gone")
21:06:37abstrakthere is a SDK or ARM tools/package availiable?
21:07:52 Nick Bagder_ is now known as Bagder (
21:07:56 Quit ilgufo (Remote closed the connection)
21:08:11abstrakarm disassembler?
21:08:16linuxstb_Download the Rockbox source code and run the tools/ script - that installs the GNU binutils and gcc for ARM
21:08:22 Join ilgufo [0] (
21:10:19abstrakall utils works under cygwin ?
21:10:46lostlogicabstrak: they should, but slow.
21:11:07amiconnFor cygwin there are packages with pre-built crosscompilers available
21:11:22linuxstb_preglow: The Samsung K5 looks interesting - it's the one with a speaker... should work on cygwin too, but it'll take hours
21:11:38amiconn(my estimation: >4 hours)
21:11:55 Join nanok [0] (n=nanok@
21:12:03preglowlinuxstb_: oooh, it's an arm5 core!
21:12:05preglow77x, that is
21:12:07lostlogicabstrak: you really might consider vmware, I think there's still a vmware image out there that's good for developing. It's apparently much quicker than cygwin to compile on
21:12:13preglowi've never worked with one of those bastards
21:12:16linuxstb_preglow: I've been saying that for weeks ;)
21:12:31*preglow turns on his attention brain centre
21:12:55linuxstb_amiconn: The cygwin packages won't have the armv5 multilibs patch.
21:13:00nanokit seems the issue with playback freezing on some files (it seemed to me ogg, but now it apears not only ogg, mp3 also) is still present
21:13:11nanokallthough not so often anymore
21:13:18amiconnlinuxstb: Not yet, no. Does current build them?
21:13:27linuxstb_Yes, but not correctly I think.
21:13:33lostlogicnanok: how often, after how long of playback, on what target?
21:14:00lostlogicnanok: does it always freeze on the same files but not on others, or are there other conditions that impact the freeze
21:14:07nanoksansa e200, i would say it happened less than 10 times today
21:14:15linuxstb_amiconn: i.e. it only works for my specific CPU, not other armv5. kkurbjun was going to investigate, and it's on my list as well. We need to fix that before updating packages.
21:14:17nanoktime of playback varies
21:14:37nanoklostlogic: that's the tricky part: i cant make out a pattern
21:14:42amiconnlinuxstb: The vmware image also doesn't have the arm multilib stuff
21:14:49lostlogicnanok: I ask you as I realize that my ipod video that's been siting here playing all day is just now frozen.
21:14:51amiconn(neither the older one nor mine)
21:14:56abstraki'll use my linux distrib...
21:15:03lostlogicnanok: it's most likely that the engine hits a need to rebuffer and somehow doesn't.
21:15:15Bagderthe tcc771 doesn't have internal sdram, but which does?
21:15:16 Quit ilgufo (Remote closed the connection)
21:15:19nanokfirst it seemed to happen on files which were ogg, but not allways. only a reboot would bring it back (after it happened nothing will play, allthough it was not crashed, menu's and the like worked)
21:15:25Bagderit has 64K sram
21:15:41linuxstb_Bagder: The TCC773 (in my device) has 2MB SDRAM builtin.
21:15:50nanoknow it also happens with mp3's, in the middle of playback (not as previously, only when starting playback of an ogg)
21:15:51lostlogicnanok: yeah, that's about how mine is I think... I'll ask you more if I need more info... it is looking like the buffering engine isn't freeing buffers properly somehow.
21:16:10nanoklostlogic: aham, i get it
21:16:32lostlogichmm... are you sure that you need to reboot to get playback going again?
21:16:41lostlogicI'm able to get it back just by stopping and then playing again.
21:16:43nanokbtw, i saw in the manual there should be a setting for the "second buffer" (anti skip), i can;t find it, was it disabled for solid-state targets?
21:16:50 Join ilgufo [0] (
21:16:57Bagderlinuxstb: you found any sensible "overview" or similar of the various tcc chips that details stuff like this?
21:17:20nanoklostlogic: when it happens at beggining of playback of an ogg, i have no other way. even switching to another file (and format) doesn;t help
21:17:21lostlogicnanok: unsure, it's under general settings|playback|anti-skip-buffer
21:17:34linuxstb_Bagder: No
21:17:48nanokyeah, that's where i looked. i don;t have that entry in the menu.. strange
21:17:52lostlogicnanok: ok, is that bug still happening at all? I think that should be gone. The one with the middle of playback is different.
21:18:12 Quit ShadowCode ()
21:18:17bertrikpcmbufdesc = 0 when it happens on my sansa
21:18:23nanokthe one with ogg happened again today, but i think only once (so definetly not so often)
21:18:39nanokthe one with mp3 seems, indeed, a whole different matter, sorry for the mixup
21:18:42lostlogicbertrik: yeah, both buffers were empty when I got it just now on my ipv
21:18:55nanoks/mp3/middle of playback
21:19:14lostlogicnanok: ok, I'm interested in more info on the beginning of an ogg one. the other one I think we have enough detail to look for.
21:19:28nanoklostlogic: what can i do?
21:19:59lostlogicnanok: try to make it happen again and see if there's something particular about the scenario where it does happen versus when it doesn't.
21:20:59nanoklostlogic: unfortunately i only had the player for a few days (rockboxed since the first half an hour or so, ofcourse ;) ), and that issue popped up a few times, but i can;t make out a pattern yet
21:21:03pixelmananok: yes, the anti-skip setting was disabled for flash based targets
21:21:24nanokafter the reboot, iirc, all is fine, and it keeps on working fine for a long time (maybe till next reboot?.. hmm)
21:21:47 Quit nicktastic ("Leaving")
21:21:47nanokpixelma: okay, i thought so ;). it makes sense, after all
21:22:27nanoklostlogic: is there any way i can make it log to a file, so i can send you some info after this happens?
21:22:31 Quit amiconn (Nick collision from services.)
21:22:35 Join amiconn [0] (
21:22:48pixelmathough the manual should be updated accordingly...
21:23:04nanoki currently have about 1G free on the player, and i don't mind keeping some free for a while, if i can help
21:23:06lostlogicnanok: you could make a logf build with buffering and playback logf t urned on, but tbh I'm not sure how to do that these days
21:23:33preglowadvanced build
21:23:43nanokpixelma: speaking of the manual, the recording options, trigger functions, are missing. and i think they are common to all targets (?)
21:23:58lostlogicpreglow: isn't there some #define stuff that needs done now too since Nico_P turned it off by default or something?
21:24:12nanoklostlogic: if you're not sure, me even less
21:24:14bertrikI won't mind to also help out to debug
21:24:26nanoklostlogic: i will try to go on that reboot thing
21:24:37nanoktomorrow, i will try to find some time to play a bit with it
21:25:17nanoklostlogic: btw, there's a "debug" setting somewhre in the menu (labeled in brakets "keep out!" :) )
21:25:28nanokwould that be any help in this case?
21:25:49pixelmananok: they are common to all targets that can record - and yes, the manual has weak points. The trigger options are best explained in the wiki IIRC.
21:27:14nanokpixelma: okay. maybe one of these days i can find some time to give a hand with the documentation. i know reading the manual convinced me to get the sansa, to rockbox it, i think it might be one of the main entrypoints for many people (about the features and such of rockbox)
21:27:27lostlogicnanok: yeah, the debug|buffering thread info for when the ogg bug happens would be useful.
21:28:12nanoklostlogic: okay, i will try to search in the docs how i can use that, and see if my 1G free is enough for it's verbosity :)
21:29:33pixelmananok: help is very welcome :) (most developers only implement new features and don't help documenting them for the average user in the manual ;P )
21:29:38lostlogicnanok: the bufferint thread is just a gui display −− logf is the one that would write to the disk and even that only stores a few kilobytes of data when you ask it to after stuff blows up
21:29:51 Quit markun (Remote closed the connection)
21:32:25nanokpixelma: i know, i am a sysadmin, so i know how "boring" writing docs is. i also know how essential it is ;)
21:32:59nanokpixelma: i also know developers/sysadmins should never be bored, at least not oto much, or they become inefficient ;)
21:33:00 Quit JRoT (Read error: 104 (Connection reset by peer))
21:33:26nanoklostlogic: aahm, i see
21:33:55amiconnlostlogic: Frozen in what way (your G5)?
21:33:57nanoklostlogic: so basically it just buffers some debug info, and it "cycles" it, and only flushes if asked to
21:34:12amiconnJust frozen gui with music still playing and backlight reacting to buttons?
21:35:54lostlogicnanok: yes, that's how logf works
21:36:25amiconnHmm, Nivo_p isn't around ...
21:36:31lostlogicamiconn: no, actually gui responsive, just music stopped and it appears that the buffering system had gotten into a situation where it thought that the whole buffer was allocated even though there was actually no real data on it
21:36:38lostlogicamiconn: yeah, he had to run
21:36:48 Quit alienbiker99 ("( :: NoNameScript 4.02 :: )")
21:37:21 Join crossfade [0] (i=4b151979@gateway/web/cgi-irc/
21:37:23amiconnI checked the wps behaviour wrt track changes more closely (while waiting for another bug to show up)
21:37:42crossfadehi is anyone here familiar with the ipod nano v1.1.3
21:37:45amiconnThe current track metadata behaves correctly, but the playlist position advances too early
21:38:39amiconnAnd sometimes the progress bar briefly jumps back to the start, but I didn't find a pattern yet
21:39:20crossfadecause im not sure how to install it onto the ipod v1.1.3
21:40:26lostlogiccrossfade: waht is 1.1.3? the firmware version? firmware version doesn't matter to rockbox installation
21:40:33 Quit ilgufo (Remote closed the connection)
21:40:39 Quit mirak ("Ex-Chat")
21:40:42 Quit mf0102 ("Verlassend")
21:40:43crossfadewell its the ipod nano mini i think its a second gen
21:40:43lostlogiccrossfade: is it the original nano or the second generation nano or the third generation?
21:40:52lostlogicrockbox doesn't support the nano 2g
21:40:58crossfadecause like when i also tried to install linux on it it did not detect my ipod
21:41:02 Join ilgufo [0] (
21:41:25crossfadeits the four gig nano
21:41:38crossfadesecond gen
21:41:46lostlogicright, we only support the nano first gen.
21:42:30crossfadeyuup its a second generation
21:42:33crossfadei looked it up
21:42:56amiconnWhy do spurious bugs never show up when you're waiting for them???
21:42:56crossfadelinux i suppose u r familiar with linux?
21:43:09krazykitcrossfade, please read the Ircguidelines
21:43:16crossfadesorry my bad
21:43:23lostlogicamiconn: observation changes the heisenbug ;)
21:43:53linuxstb_crossfade: If you're about to ask about ipodlinux, it works on the same ipods Rockbox does, no more.
21:43:59crossfadebut back to rockbox cause id really like to mod my ipod second generation but like i don't know how to install it
21:44:24lostlogiccrossfade: as I've said repeatedly, rockbox does not work on the 2nd gen ipod nano.
21:44:30crossfadeo k
21:44:34linuxstb_Rockbox doesn't work on the 2nd gen Nano - it's a completely different device internally to other ipods, and no-one has ported Rockbox to it.
21:44:45 Join markun_ [0] (n=markun@rockbox/developer/markun)
21:46:06crossfadeso i presume that goes for the linux osx as well
21:46:16crossfadedamn o well i was hoping it could thanks for the help
21:48:04 Quit crossfade ("CGI:IRC (EOF)")
21:49:04 Quit donutman25 ("ChatZilla [Firefox]")
21:52:17xyshi, i recently started using rockbox on my ipod 5g 30gb, is there anything that can be done to increase battery life?
21:53:12lostlogicxys: use less features (ie eq, replaygain, crossfade, crossfeed, peakmeter)
21:54:04lostlogicxys: or help us figure out what might be using extra battery and/or get rockbox actually running on both cores to reduce the cpu speed needed
21:54:59amiconnThe core clock isn't that much of an issue
21:55:02xysi was looking at the spec sheet for the 5022 and was wondering if you guys were using both cores
21:55:26xysis there work on that?
21:55:29amiconnThere are 3 main power suckers atm: (1) we need to find out what internal modules we can switch off, and how
21:55:44linuxstb_xys: A little. e.g. mpegplayer uses them, and there is work being done to make core Rockbox dual-core safe.
21:55:45lostlogicxys: thanks to jhMikes and safetydan both cores are runnable, but currently nothing runs on the second core.
21:55:56lostlogicoh, forgot about mpegplayer.
21:55:58amiconn(2) (G5 only) We need to figure out backlight brightness. Full brightness is a real power sucker
21:56:09amiconn(3) We need to figure out disk poweroff
21:56:27linuxstb_lostlogic: Also, wasn't it dan_a who did the initial dual-core work, not safetydan?
21:56:34lostlogicamiconn: we don't currently turn off the interface when we spin down?
21:56:40amiconnOn G1..G3 where we did (1) and (3), we get runtimes on par, or exceeding, OF runtime
21:56:49lostlogiclinuxstb_: yeah, sigh, I've been away too long ;)
21:57:05xysis spindown different from disk poweroff?
21:57:06amiconnAnd on G1..G3 rockbox runs single core
21:57:11linuxstb_lostlogic: You have - but welcome back ;)
21:57:16amiconnxys: yes
21:57:56lostlogiclinuxstb_: thanks :-P I still might turn out to be an illusion
21:58:07amiconnWhen spun down, all HDDs still need a bit of power for their ata interface. Most, if not all, hdd based mp3 players have an electronic switch to cut the power to the hdd when not in use
21:58:23amiconnOn the ipods, except G1..G3, we don't know how to do that yet
21:59:14xysI was wondering the legality of disassembling the ipod firmware also
22:00:46linuxstb_xys: I think that depends on the country you are in.
22:01:04 Nick Buschel_off is now known as Buschel (
22:01:42Buschelamiconn: maybe you're right with assumption that the x31 starts transmission to display
22:02:14amiconnBuschel: 0x31 is a general 'go' command, also for the other instructions
22:02:27Buschelamiconn: if you send x31 rigth after setup_rect the fps goes up to 36.6fps (@30MHz).
22:02:39amiconnI tried all 4 I found, both individually and in combination like the rom does
22:02:56Buschelamiconn: but you'll display the last screen screens update then
22:03:00amiconn3 seem to do just nothing, one is a reset code
22:03:43amiconnSending 0xfff70008 to 0x1f8 followed by 0x31 to control produces a white flash
22:04:32amiconnBuschel: Did you try to not sending the write address everytime, e.g. in setup_rect?
22:04:43Buschelamiconn: we may use the 14ms of the display update for the routines which copy data to lcd data port
22:04:50amiconnI suspect that the address pointer auto-increments
22:04:52 Join Domonoky_ [0] (
22:05:43 Quit darkless (Read error: 110 (Connection timed out))
22:05:49amiconnBuschel: Much simpler - we can let lcd_update_rect and lcd_yuv_blit just let write to the appropriate addresses, and then set a flag 'update needed'
22:06:08Buschelamiconn: which one? the E0020?
22:06:17amiconnThe scroll thread could take care of that, i.e. check whether the bcm is ready for update, and then send 0x31
22:06:35amiconnI think 0xe0020 is just the start addres
22:06:44 Quit abstrak ()
22:07:00nanoklostlogic: i will be out in a minute. i will try to update to the latest build (clean, with no themes and such except for the default pack), and if it happens again, i will let you know (the ogg freeze problem), and if it does, maybe we can talk about some "serious" debugging/logging
22:07:10amiconnWe could set up a permanent fullscreen rectangle, and then render into the appropriate region
22:07:15nanoklostlogic: does that sound good?
22:07:25lostlogicnanok: sounds good, also nico_p is your man for such things if I'm not around ;0
22:07:48nanoklostlogic: my current install is the build from 27 btw
22:07:52 Part przemhb
22:08:00nanoklostlogic: okay, good to know
22:08:02bertriknanok, can you reproduce FS #8037?
22:08:23amiconnBuschel: I mean, the first word of the data we're transferring is written to 0xe0020, the next one to 0xe0022 etc
22:08:24nanokbertrik: uhm, let me check, i don;t know the number :)
22:08:24lostlogicnanok: ah, the ogg freeze might have been picked up by one of Nico or my commits since then.
22:09:14nanoklostlogic: i updated to that one because someone here (maybe amiconn? not sure) said it was fixed (at that time just 8 hours old build or so)
22:09:32nanoklostlogic: but i am laos thinking maybe, somehow, i didn;t quite "catch" it
22:09:36amiconnThat's what I mean with the autoincrementing addresses - it should also work in setup_rect that we just set the address to 0xe0000 and then transfer all parameters in succession
22:09:52nanokdamn ;
22:10:00nanokit allways happen when i smoke and type..
22:10:06amiconnI'll check that theory, after my experiments wrt the lcd freeze
22:10:31lostlogicnanok: we've been making lots of improvements to the buffering stuff since Nico first did the big MoB commit last week.
22:11:17amiconnlostlogic: Do you have an idea how to fix the too early advancing playlist position?
22:11:29nanoklostlogic: cool. does this include some possible power consumption improvements, by any chance (i know with the sansa that's not where it is lacking, but still .. :) )
22:11:39 Join petur [0] (
22:11:45lostlogicamiconn: use the pcmbuf end of track callback to advance it instead of the buffer over... but I'm not sure of the full implications of that
22:12:01lostlogicnanok: no.
22:12:07amiconnThe sansa also has the power consumption issues, just not as significant (the hdd power issue doesn't apply to itz)
22:13:03nanokbertrik: aaaahm. sounds like my litle devil (FS #8037), i will try it also
22:13:45nanoklostlogic: :). okay, you can't blame a man for hoping.. :P
22:14:50 Quit ender` (Read error: 104 (Connection reset by peer))
22:14:58nanoki understand the power management is a work in progress on the sansa (don;t know how it goes though, i hear quite sllow..). i did notice though that it is not as bad as it seems
22:15:08 Join ender` [0] (
22:15:09nanokchanging the theme and disabling the eq did wonders
22:15:16 Join webguest50 [0] (i=4a469ecd@gateway/web/cgi-irc/
22:15:27lostlogicnanok: yeah, often just disabling the 'extras' makes a big improvement.
22:15:44bertriknanok: the theme influences power consumption?
22:15:46nanoki think that should be documented somewhere also (i found it in a mail thread, quite difficult to get ot it), as in "power saving tips"
22:16:26lostlogicnanok: you're welcome to update our manual and submit a patch :-D
22:16:29nanokbertrik: well, a very "heavy" theme seems to be able to drain quite a lot. i suspect that also depends on the target though (the dev's can tell us i think)
22:17:07nanoklostlogic: as i said earlyer, i think i will try to help with that, as i am no coder, not much more i can do, though
22:17:07lostlogicbertrik: peakmeter and progressbar specifically do consume cpu cycles and cause more frequent LCD updates.
22:17:10amiconnBuschel: If we do it clever enough, we could try to send the 0x31 at exactly the right moment so that we're running just a tiny bit ahead of the bcm's internal processing
22:17:25lostlogicnanok: if you hang out long enough, you might be surprised ;)
22:17:37 Quit merbanan ("Leaving")
22:17:58amiconnMeh, heisenbug ;\
22:18:10webguest50i have just converted my e200r into a e200 and i am dual-booting with rockbox. For some reason, when i connect it to the computer, nothing happens. it says connected on the sansa screen and when i plugged it in it said found new hardware. The neccessary files were not foundand nothing appears in my computer. When i used e200r, it connected fine but now it won't. Any suggestions. (i'm using sansa e200)
22:18:33***Saving seen data "./dancer.seen"
22:18:45nanoklostlogic: i don't hang out. i need music like i need nicotine :), and i allways wanted an opensource firmware (i;m fed up with all the limitations, praying and waiting). so if i'm in, i'm in. it all comes down to how much time i can spare, from this point on ;)
22:19:01Domonoky_webguest50: use the OF firmware for usb..
22:19:22lostlogicnanok: that's kinda what I meant by hang out −− if you're here often and start submitting patches for the manual, you just might find yourself learning to code too.
22:19:43nanokbut right now, i really have to watch this movie with my girlfriend, or i won;t be breathing when i wake up tomorrow morning (so i won't be much help for anything, in that case)
22:20:29nanoklostlogic: oooh, i doubt about coding. anything beyond some bash or perl is way out of my league (perl is allready voodoo stuff ;) )
22:20:50nanoksee you, guys
22:20:58nanokthanks a lot ;)
22:21:38nanokwebguest50: maybe change de usb mode to "mass storage device" on the sansa?
22:21:52 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:22:00nanokahm, sorry
22:22:09 Quit Domonoky (Read error: 110 (Connection timed out))
22:22:18nanokmissunderstood. clear sign i must take a break ;)
22:22:45 Join Zagor [0] (
22:23:23 Part bluey ("Leaving")
22:23:34webguest50nanok: It is on MSD mode
22:23:42nanok(or maybe not, i think rockbox never ays "connected", sounds like the OF)
22:23:53nanokwebguest50: uhm, i never can remeber those acronyms
22:23:57nanoktry the other one :)
22:24:05nanokwhat the hell is that one called?
22:24:30webguest50i tried both MSC and MTP modes
22:24:37scorche|wyou want MSC mode
22:24:59nanokwebguest50: windows? or something else?
22:25:26webguest50MSC connects but when i try to install it after the "found new hardware" window pops up, it doesn't find the files it needs i think
22:25:57nanokwebguest50: i think scorche|w is right, btw, msc is the one i am using also
22:26:12nanokyou try to install what exactly?
22:26:40webguest50i already had rockbox installed and rockbox works fine
22:26:57Domonoky_i think webguest just have to boot into the sansa firmware to get usb working.. rockbox will present its usb as "unkown device" :-)
22:27:00nanokdo you see a drive letter for for the player? (it is windows i assume, right?)
22:27:25 Quit ilgufo ("So Long, and Thanks For All the Fish -")
22:27:29Domonoky_webguest50: you are in the sansa firmware when you connect usb ?
22:27:36nanokDomonoky_: i have a hunch he allready is, rockbox in my experience never says "connected" on the sansa
22:27:38webguest50hold on, i'll give you exactly what it says (it said Removable E when i used e200r but now i don't see anything
22:27:39bertrikwindows wouldn't recognise my sansa until I installed the software from the cd-rom
22:27:49webguest50domonoky: yes
22:27:58 Quit Echel0n (Connection timed out)
22:28:20nanokbertrik: oh really? weird, i assumed it should work out of the box with msc mode at least
22:28:32nanokit does on linux :-D
22:28:57bertrikI'll wipe the registry entries and try again to be sure
22:29:19 Quit Frazz ("Leaving")
22:29:21nanokwebguest50: try to check if there is a new letter, like e or something, anyway, even if windows doesn't tell you there is anymore (maybe it is just quiet)
22:30:15webguest50Windows has found the following new hardware:
22:30:24webguest50sorry, did that ny accident
22:32:29 Quit nanok ("fourtytwo...")
22:33:45 Quit J3TC- (Read error: 110 (Connection timed out))
22:33:59webguest50it asks if i want to automatically search for a better driver or to specify the location of the driver. I selected automatic. Then it says windows was unable to locate the software for this device.
22:35:47 Quit Alonea ("ChatZilla [Firefox]")
22:37:12 Quit xys ("Ex-Chat")
22:38:49 Join J3TC- [0] (
22:43:19Bagderit seems we never got any scans/fotos of a gmini
22:43:50Bagderthere are photos of an xclef mt-500, claimed to be very similar to the gmini
22:44:07 Join Echel0n [0] (
22:44:43 Join ender1 [0] (
22:45:05 Quit ender` (Read error: 104 (Connection reset by peer))
22:47:21Bagderthe telechip page grew a lot... :-)
22:48:00 Join markun [0] (n=markun@rockbox/developer/markun)
22:48:37 Quit markun_ (Read error: 104 (Connection reset by peer))
22:48:53bertrikis there a guide somewhere on how to use debug and logging with rockbox?
22:49:13 Join safetydan [0] (i=dc9d468b@rockbox/developer/safetydan)
22:50:59 Quit webguest50 ("CGI:IRC (EOF)")
22:53:38bertrikI compiled for debug and logf, but I can't notice any difference with a normal build
22:53:44 Nick ender1 is now known as ender` (
22:53:57bertrikROCKBOX_HAS_LOGF is still #undef'ed in autoconf.h
22:54:44preglowamiconn: here?
22:54:52Zagorbertrik: the behaviour recently changed. you now have to add LOGF_ENABLE to those files you want logging enabled in
22:55:13bertrikok thanks
22:55:28Zagorotherwise the log is quickly filled with stuff you're probably not interested in
22:55:34lostlogicbertrik: so if you're looking at the playback and buffering bugs, you'd add #define LOGF_ENABLE to the top of playback.c and buffering.c
22:56:22amiconnpreglow: yes
22:56:23Zagorand iirc you need to add the define before #include "logf.h"
22:56:46bertrikZagor: yes, I enabled that one
22:57:04bertrikI should also probably enable the ROCKBOX_HAS_LOGF in autoconf.h?
22:57:19preglowamiconn: look at clip_sample_16() in dsp.c, the fast clip trick i used in filters_cf.S for speex should be way faster than that, right?
22:57:29bertrik(i'll know if it worked in 15 minutes)
22:57:43Zagorbertrik: that should have been done by tools/configure
22:58:42bertrikI can just type "DL"<enter> at the advanced options question of tool/configure?
22:59:37ZagorI don't know actually, I've only ever used either option
22:59:43bertrikIf i just type "DL"<enter>, the ROCKBOX_HAS_LOGF is #undef'ed in autoconf.h
23:00:14 Quit Domonoky_ ("Trillian (")
23:00:26amiconnpreglow: I'm not sure. This one should be pretty fast too
23:00:27 Join Domonoky [0] (
23:00:39amiconnYou can speed up the clipping case a bit with asm
23:00:40preglowthe clipping in sample_output_dithered is really broken for frac_bits >= 30...
23:00:42 Join webguest84 [0] (i=5439acbf@gateway/web/cgi-irc/
23:00:59Zagorbertrik: ok seems that doesn't work then. in that case yes enable ROCKBOX_HAS_LOGF manually in autoconf.h
23:01:17Bagderbertrik: you'll have to select d and l individually
23:01:28amiconnThe non-clipping case should in fact be faster than what you do in speex, but it can't be used for arbitrary ranges
23:01:29Bagdernot both at the same time
23:01:45Bagdertype d[enter] then L[enter]
23:02:22preglowamiconn: nah, i figured that
23:03:27amiconnIn speex you need a non-standard range. The method in clip_sample_16() only works fast for clipping to 16 bit (and 8 bit)
23:04:10 Quit webguest84 (Client Quit)
23:04:38Buschelamiconn: gotta leave, maybe you've got the breaking through idea? I need to let the news settle a bit :o)
23:05:21amiconnIn cf asm I would write it like this (clipping %d0): move.l %d0, %d1; ext.l %d1, cmp.l %d0, %d1; beq.s 1f; swap %d1; clr.l %d0; move.w %d1, %d0; 1:
23:06:01 Quit Buschel ()
23:06:16amiconnThat would be 6 cycles non-clipping, and 7 cycles clipping
23:06:52 Quit TotallyInfected ()
23:06:55 Join midgey [0] (
23:06:58amiconnHmm, forget that
23:07:18*amiconn silly
23:09:25 Join webguest91 [0] (i=53c7a3b9@gateway/web/cgi-irc/
23:09:59bertrikargh, won't compile, errors in kernel.c
23:10:00 Join JRoT [0] (
23:10:14 Part JRoT
23:10:17 Quit jgarvey ("Leaving")
23:10:18 Quit webguest91 (Client Quit)
23:12:06preglowman, that clipping in sample_output_dithered looks weird
23:12:59preglowman, that is one bugged piece of shit
23:17:40 Quit Llorean ("Leaving.")
23:19:25amiconnYay! My theory is correct
23:20:06amiconnThe bcm auto-increments the last used write address, so lcd_bcm_setup_rect() needs to do a lot less
23:20:45*amiconn wonders about the meaning of the 'cmd'
23:20:55bertrikI think I cannot use DEBUG at all with a sansa....
23:21:17Bagderbertrik: what do you want to achieve?
23:21:35bertriklook into the playback.c problems
23:21:51Bagdersurely logf is what you want then?
23:22:00Bagderunless you go for a sim build
23:22:11lostlogicbertrik: you probably just want logf, debugf is for serial debugging or sim
23:22:16bertrikI'll put up a patch for kernel.c debugging
23:22:57bertrikok thanks
23:24:50 Quit Domonoky ("Trillian (")
23:26:09 Join freqmod__nx [0] (
23:26:39 Quit freqmod__nx (Remote closed the connection)
23:30:08 Quit ompaul (Client Quit)
23:31:03preglowamiconn: can you think of any fast ways of doing clip_sample() that doesn't mess up with extreme values of min/range? it messes up if range is 2^31 - 1
23:31:19 Quit jhMikeS (Nick collision from services.)
23:31:25 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:31:56 Join midgey_ [0] (
23:35:38 Quit J3TC- (Read error: 110 (Connection timed out))
23:38:13 Join TMM [0] (
23:47:06 Quit tpickers66 (Read error: 113 (No route to host))
23:47:44 Quit midgey (Read error: 110 (Connection timed out))
23:52:33 Quit midgey_ ()
23:55:14 Quit ender` (" To get as fewest unhappy people as possible, always bully the same ones.")
23:58:00 Quit MethoS-- (Remote closed the connection)

Previous day | Next day