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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2007-11-27

00:00:00XavierGrCongratulations Bagder \o/
00:00:16 Join zajacattack [0] (i=42617b7b@gateway/web/cgi-irc/
00:00:29pixelmarasher: ah, in German it's only seperate up to 12 as in quite a few other languages
00:00:54zajacattackso, how is the work on the USB stack going?
00:02:06peturBagder: congrats!
00:02:39*scorche|w cheers for Bagder
00:03:30*rasher adds yet another half-baked patch to the tracker
00:03:38zajacattackyes, congrats Bagder
00:03:44rasherWith any luck, someone will take one and run with it some day
00:06:32 Join Traveler2 [0] (
00:06:44Traveler2yay, got my first plugin working
00:07:11Traveler2all it does is draw a rectangle and wait till you press the quit button, but its a start
00:07:13preglowwhat's the price? free software? :P
00:07:29Bagderhaha, who knows
00:11:36Nico_PBagder: Wear a rockbox shirt at the ceremony ;)
00:11:42 Join J3TC- [0] (
00:12:41BagderNico_P: I won't even make it to the ceremony! :-/ I'm doing a speech at that conference, but the day after the ceremony as I'm going there and back again on the same day (~3 hours train trip each way)
00:13:52Nico_PBagder: ooh, shame... what if you get the award?
00:14:10 Join axion [0] (
00:14:20BagderI don't think that'll happen, but I won't be there anyway
00:15:18zajacattackBagder: well, good luck on your speech
00:16:22BagderI might wear my rockbox shirt then ;-)
00:17:09markunthe white one?
00:17:27pixelmamight be as fun as to see the rockbox shirt in China, markun found the pic I believe :)
00:17:33markundo you also have curl t-shirts?
00:17:49Bagdermarkun: nope, we don't
00:18:10*linuxstb notices the Gigabeat S has 64MB of RAM, not 32MB as configure says...
00:18:18Nico_Pnice :)
00:19:24preglowamiconn: you should have removed those extreme fade out times too, heh
00:20:03 Quit axion ("leaving")
00:20:13amiconnOverlooked one get_talk_id function that needs adjustment
00:20:16preglowamiconn: player warnings, of all things
00:20:16 Join axion [0] (
00:20:44 Nick axion is now known as axionix (
00:20:51*preglow finally sets fade in time to 200 ms
00:20:59amiconnYeah jumpscroll_getlang(). Don't even need to look at the log
00:21:04 Quit axionix (Client Quit)
00:21:16 Join axion [0] (
00:21:21preglowand a cute little green delta too
00:21:56 Quit axion (Client Quit)
00:22:09 Join axionix [0] (
00:22:18amiconnI hoped for that green delta :)
00:22:53 Quit axionix (Client Quit)
00:22:54 Quit Billenium ("Ex-Chat")
00:23:05 Join axionix [0] (
00:23:44amiconnHrrm, that reminds me. This jump scroll thingy needs to be either repaired+extended, or removed
00:23:54*amiconn would prefer the former
00:23:55 Join donutman25 [0] (n=chatzill@
00:24:21preglowamiconn: btw, buschell stripped everything but the vuv change from his patch, commit now?
00:24:34*preglow doesn't like having obvious commital candidates hanging around in the tracker
00:24:43amiconnYeah I know. Was busy with that backlight fading stuff
00:25:01 Quit Zagor ("Client exiting")
00:26:54rasherAny objections to committing the matrix demo? Someone fixed my awful coding.
00:28:11preglowgo ahead
00:28:15preglowhow many targets does it support?
00:28:28*amiconn builds for G5
00:28:57rasherI think it works on all colour targets, except for those lacking keymaps
00:29:10linuxstbWhat keys are needed?
00:29:21rasherjust speed up/down and exit
00:29:34rasherSo it's not like that's a huge problem
00:29:37Nico_Plinuxstb: dual boot works fine :)
00:29:50rasherThe thing's here by the way:
00:30:04rasherHe also created one that asks for a password when exiting, but I'm not sure about committing that
00:30:16stripwaxDid anyone get a chance to try on nano?
00:30:26linuxstbNico_P: Nice... I think I would still like to get it working better before committing though - i.e. having the dual-boot code in the Rockbox bootloader, rather than hard-coded in mknkboot.
00:30:48preglowrasher: drop the password stuff, none of the other "screensavers" do it
00:31:01preglowrasher: and if we ever do anything like that, it should be just one plugin or one core feature
00:31:11*stripwax hmms
00:31:23rasherpreglow: Agreed.
00:31:25Traveler2whats this about a dual boot code?
00:31:27stripwaxSliding Puzzle shows up under "Applications". Should that be Games?
00:31:29 Nick midkay_ is now known as midkay (
00:31:33preglowstripwax: deed
00:31:56stripwaxpreglow - ok, will add that to my patch too
00:32:08preglowstripwax: what patch?
00:32:13safetydanamiconn: is your next step to move the sound units out of firmware? I was most of the way through that last night, but the recording string held me up.
00:32:21stripwaxpreglow - I'm updating the sliding_puzzle bitmap patch
00:32:45stripwaxpreglow - that is to say, I'm working on it at the moment, I haven't updated the tracker yet
00:32:49amiconnsafetydan: I am planning to do some greyscale work next
00:33:18amiconnToo much high-level work in succession isn't good for my motivation ;)
00:33:41safetydanyou sound like preglow complaining about doing ui code :)
00:33:58linuxstbrasher: Do you have an ipod?
00:34:00preglowstripwax: ahh, ok
00:34:01amiconnYeah, low-level code is definitely more fun
00:34:04rasherlinuxstb: nope - sansa
00:34:27Traveler2I got an ipod, planning to upgrade to sansa e200
00:34:37 Quit lee-qid (Read error: 110 (Connection timed out))
00:34:43 Join lucas42 [0] (
00:34:44preglowsafetydan: about ui code, did you ever flesh out that new eq screen any more?
00:34:47linuxstbrasher: I've looked briefly at matrix, and it doesn't look like it will increase/decrease speed on repeat events - which is annoying with the wheel. Also, it seems odd to use the wheel on the ipod, but not on the Sansa.
00:35:14rasherlinuxstb: isn't that basically a case of adding |BUTTON_REPEAT somewhere appropriate?
00:35:44zajacattackhey, has anyone here been working on the new USB stack?
00:35:48amiconnIs there any reason why matrix wouldn't work on greyscale/mono?
00:36:07rasheramiconn: not really. It'll look silly on mono though, I predict
00:36:11Bagderzajacattack: zagor's done most of it and he left a short while ago
00:36:37linuxstbrasher: I think so...
00:36:50 Join Inkey5 [0] (n=james@
00:36:59linuxstbrasher: How many speed settings are there? IMO the wheel only makes sense with a large range of values.
00:37:11Inkey5how do I transfer files to my sansa e250 w/ rockbox on it?
00:37:23Inkey5windows does not recognize a storage device
00:37:26*linuxstb reads the code and it seems there is no upper limit to how slow it gets...
00:37:29Inkey5osx doesnt either
00:37:36rasherlinuxstb: iirc, it just adjusts a sleep, so in theory limitless
00:37:47krazykitInkey5, that'd be because you need to boot to the OF for usb transfer right now
00:38:16zajacattackso, why can some older devices do USB but not the new ones?
00:38:25 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
00:38:30stripwaxInkey5 - the usb code is under development and won't do the right thing if you connect your device to your PC while rockbox is running on your device
00:38:45rasherzajacattack: the old models use a hardware usb bridge
00:38:57zajacattackand what do the new use?
00:39:02Inkey5reboot and hold left?
00:39:07linuxstbrasher: At the very least, I think the missing kemaps need to be added for all color targets. Looking at it, I think that's just the H10 and C200?
00:39:13krazykitInkey5, yep
00:39:14zajacattacko, i see
00:39:43rasherlinuxstb: I'll fix that and the wheel thing and commit?
00:39:46 Quit Thundercloud (Remote closed the connection)
00:40:08Inkey5is there an etr on the usb thing?
00:40:21Inkey5can we expect a fix soon?
00:40:26 Join Thundercloud [0] (
00:40:47rasherInkey5: it's not a fix. More along the lines of "writing a complete driver and usb stack"
00:40:50n1sInkey5: we don't estimate dates ot times for anything
00:40:52Inkey5well, right
00:41:01Inkey5thats cool
00:41:11Inkey5I appreciate what you guys have done, nonetheless
00:41:16rasherSome work has been done, and the rest could take days, or months. Or never work at all.
00:41:24linuxstbrasher: It looks to me as if it will work on greyscale targets - have you tried compiling it? Also, does it make sense for all LCDs to use the same sized bitmaps? (sorry for being fussy....)
00:43:17*jhMikeS notices the sansa bootloader now boots rockbox by default even when plugged
00:43:38Lloreanrasher: You could simply use the actions system for the keymaps.
00:44:10rasherLlorean: I can never remember if the wheel will do the right thing.. I'm guessing not.
00:44:32LloreanYeah, it wouldn't likely use the wheel
00:44:50linuxstbDoesn't bubbles use the action code? That's using the wheel.
00:44:58LloreanOr rather, it'd invert the wheel
00:45:30LloreanWell, maybe.
00:45:34pixelmalinuxstb: the e200 has an exception in bubbles...
00:46:05LloreanActually, looks like the ACTION_STD context would work well
00:46:11linuxstbpixelma: Why? Isn't it comparable to the ipod?
00:46:19LloreanScroll down, and "Down with the stick" for h120 are equivalent
00:46:32*linuxstb regrets asking pixelma about plugin actions ;)
00:46:52pixelmalinuxstb: probably because someone was afraid to break other plugins
00:47:21pixelma(JdGordon did that commit himself...)
00:47:44linuxstbShouldn't matrix be the same as the settings? i.e. SETTINGS_INC, SETTINGS_DEC ?
00:47:45pixelmaLlorean: plugin use their own system of button actions
00:48:00rasherI think the plugin action suffers from: a) being implemented on bubbles first b) trying to d too much c) poor documentation
00:48:01*JdGordon did what?
00:48:02Lloreanpixelma: I thought at least one plugin used the ACTION_STD mappings?
00:48:21Lloreantext editor?
00:48:38pixelmayeah because it uses the core virtual keyboard
00:48:54LloreanBut couldn't another plugin use ACTION_STD mappings too?
00:49:00pixelmaJdGordon: the e200 keymap change in bubbles
00:49:03LloreanIf it just needs generic "Next, Prev, and Cancel" keys?
00:49:36JdGordonyes, plugins can use the core mappings also
00:50:30 Quit linuxstb ("ChatZilla 0.9.79 [Firefox]")
00:51:01 Quit MethoS- (Remote closed the connection)
00:51:54 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:52:20*stripwax hmms again
00:52:43stripwaxI see from apps/plugins/CATEGORIES that sliding_puzzle is indeed flagged as Games. But when I make install it ends up in apps. Any ideas why that is?
00:52:47 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
00:52:51amiconnpreglow: Tested and committed.
00:52:58 Quit ender` (" But there, everything has its drawbacks, as the man said when his mother-in-law died, and they came down upon him for the f")
00:53:42preglowamiconn: excellent
00:53:45JdGordonstripwax: viewers.config can move it also
00:54:12stripwaxJdGordon - ah. right. That's good to know..
00:54:31stripwaxyep, that's it. thanks!
00:54:48rasherHrm, (button & (~BUTTON_REPEAT)) will just unset the button_repeat bit, right?
00:54:59stripwaxshould do
00:55:19 Quit petur ("Zzzzz")
00:55:28 Quit advcomp2019 (Read error: 110 (Connection timed out))
00:55:47 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
00:56:50Nico_Plinuxstb: the dual boot is really a plus... and I can update while in the OF without a problem and it doesn't erase my files
00:57:12 Join Lars_G [0] (n=Lars@unaffiliated/lars-g/x-000001)
00:57:19Nico_Pmaybe we can think of loading a file now
00:57:26Lars_Gapps/settings_list.c has unresolved merge at the end
00:57:30 Join aliask [0] (
00:57:35Lars_Gdah it's MY merge
00:59:24linuxstbNico_P: Glad it's useful. I'm experimenting a little now, and it seems that RAM is 64MB starting at 0x88000000 - so up to 0x8bffffff
01:00:10preglowlinuxstb: you've got a gigabeat s now?
01:00:12linuxstbIIUC, eboot.bin is being loaded to 0x88000000, and then nk.bin is being loaded at the addresses specified in the nk.bin chunk headers, which start at 0x88200000
01:00:17*amiconn will try to find out where this excessive 2KB alignment on PP502x comes from
01:00:18Lars_GHmmm is it me or the usb stack was dropped?
01:00:26linuxstbpreglow: I've had one for a while, but have been resisting the urge to hack...
01:00:36preglowlinuxstb: more than i would have been able to
01:00:56preglowamiconn: why would pp need to align by 2kb?
01:01:14preglowlinuxstb: how does one put rockbox on it, btw?
01:01:15jhMikeSgcc won't even accept alignments that large I thought
01:01:25amiconnThat's what I don't know... but the build table shows that it does
01:01:26aliasklinuxstb: Did you get dual booting working?
01:01:29jhMikeSthough the theres 16384 bytes array for usb
01:02:00 Quit scorche|w ("CGI:IRC")
01:02:01linuxstbaliask: Yes, but I'm still not completely understanding what is going on. I got it working by injecting 28 bytes of code (the hold-switch check and branches) to 0x88200000
01:02:42linuxstbaliask: Also, are you aware the S has 64MB of RAM? tools/configure states 32MB.
01:03:14aliasklinuxstb: Yes, I did. But in the configure script I noticed a comment that said "always 32" so I didn't change it.
01:03:44 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:04:34rasheryay, reboot on usb works on sansa
01:04:53 Part LinusN
01:05:43preglowlinuxstb: is it now possible to rockbox a gigabeat s with no other tools than our own?
01:05:55jhMikeSheh, after SVN up, 3g myteriously boots again :p ... to heck with it, I'm committing the DC for 5002 since the problems are obviously unrelated.
01:06:09 Part Lars_G
01:06:36preglowjhMikeS: are these the same bugs amiconn are having on his 2g?
01:07:10jhMikeSI think so
01:07:52 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
01:08:01 Quit linuxstb (Nick collision from services.)
01:08:06 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
01:08:10jhMikeS3g is happily playing SPCs again...just from SVN up
01:08:37hcsI wonder how this compares to IPL's SPC player
01:08:53preglowwell, then i guess a binary search for an offending commit is coming up, though i can see how such a search might be difficult to do accurately...
01:08:55linuxstbpreglow: Yes. Although you need a copy of the original firmware to patch, and I think the update tool will only work on Linux.
01:09:03preglowhcs: probably favourably, it is top quality
01:09:22linuxstbaliask: Have you tried creating an nk.bin with just Rockbox in it?
01:09:22hcsjhMikeS: is kickin' ass
01:09:29JdGordonaliask: that "always 32" comment is because you can chose the RAM size of some of the targets...
01:09:29preglowlinuxstb: well, sure, i hardly expected to not need the original firmware yet, heh
01:09:33 Quit gromit`` (Read error: 104 (Connection reset by peer))
01:10:06preglowthe lust for gigabeat s strikes again
01:10:13 Join gromit` [0] (
01:10:18Lloreanlinuxstb: Can we extract that copy of the original fw from the device?
01:10:27linuxstbpreglow: You _really_ don't want to see the original firmware though...
01:10:37aliasklinuxstb: As in, without the OF?
01:10:41linuxstbaliask: Yes.
01:10:49Lloreanpreglow: Yes, I have to agree with that. Don't boot the OF, and absolutely don't try to actually *use* it or navigate in it
01:10:49preglowlinuxstb: i really don't plan on it
01:10:58linuxstbLlorean: I don't think so.
01:10:59JdGordonlinuxstb: cant be worse than the irivers...? can it?
01:11:15preglowi thought iriverfw was ok
01:11:22linuxstbI've forgotton what the iriver firmware is like, thanks to Slasheri it no longer exists.
01:11:31aliasklinuxstb: I tried, but it was difficult to get an nk.bin which would be allowed to be sent by the gigabeat v updater - the nk.bin format isn't well documented
01:11:42jhMikeSbootloaders seem ok after rearranging crt0-pp-bl.S too
01:11:46preglowlinuxstb: so, what shape is the port in?
01:12:07amiconnjhMikeS: gcc won't accept it?? Look at usb-drv-pp502x.c, line 310...
01:12:17amiconn__attribute((aligned (2048)));
01:12:32linuxstbpreglow: I'm not sure really - I've just been working on getting dual-boot to work, to make installing test code easier. I haven't actually tried any code, apart from the "hello world" bootloader which is in svn.
01:12:34*amiconn wonders why the heck such an alignmend would be needed
01:12:40rasherSoooo... Who to credit for the matrix plugin? It was originally cmatrix for linux, ported to ipl/podzilla, ported to rockbox...
01:13:10preglowlinuxstb: so, rockbox isn't working yet?
01:13:31*Nico_P doesn't fine the gigabeat S OF *that* bad
01:13:32aliasklinuxstb: However, you can see my efforts in /tools/gigabeats.c (for scramble.c)
01:13:37Nico_Ppreglow: far from it
01:13:40***Saving seen data "./dancer.seen"
01:13:51preglowany major difficulties left?
01:13:59jhMikeSamiconn: hmmm....maybe the assembler was whining then
01:14:12jhMikeSThere is a limit though
01:14:31amiconn.align in gas means actually 2^n for most targets
01:15:11stripwaxsliding puzzle bmp / albumart patch updated
01:15:35zajacattackok, now, the e200 v2 have completly different HW, so there's no way to update a v1 to a v2, right?
01:16:00linuxstbzajacattack: Yes, change the hardware...
01:16:05rasherzajacattack: Not unless you get out your soldering iron..
01:16:13jhMikeS.balign is just a byte align though and has a small limit
01:16:39 Join alienbiker99 [0] (
01:17:00 Quit advcomp2019 ("Nice Scotty, now beam my clothes up too!")
01:17:06linuxstbaliask: I posted my dualboot patch here -
01:18:03aliasklinuxstb: Will test
01:18:51Traveler2so wait, rockbox will install on new sansa e200's, right?
01:19:02JdGordonstripwax: your change to viewers.config looks wrong? doesnt that put it in .rockbox/rocks instead of /rocks/games ?
01:19:02n1snot the v2 models
01:19:13LloreanIs the Gigabeat S the i.MX31 or i.MX31L?
01:19:28linuxstbWhat's the difference?
01:19:32LloreanNo 3D graphics.
01:19:33n1sno graphics accelerator
01:19:37stripwaxJdGordon hm - possibly, if I didn't upload the correct patch..
01:19:42linuxstbDamn, I want a refund...
01:20:09n1sthough it's quite overkill for an mp3 player even without accelerated open gl...
01:20:13rasherLlorean: I thought you were joking, but then again, it's the Gigabeat S..
01:20:16linuxstbSo "L" is the "lite" version? ;)
01:20:26amiconnJdGordon: Did you check out my formatter/getlang unification?
01:20:29Lloreanrasher: Considering how much overkill this chip already has...
01:20:32LloreanJAVA acceleration?
01:20:41rasherWhat, seriously?
01:20:42JdGordonamiconn: yeah, had a quick glance.. looks fine
01:20:54Lloreanrasher: "Jazelle Java acceleration"
01:21:13Traveler2so how do I know if a sansa e200 is v2 or not?
01:21:14*JdGordon wants openGL accelerated lists in rockbox!
01:21:14amiconnLlorean: Well, java definitely needs acceleration even on multi-GHz PCs to be bearable...
01:21:17linuxstbAnd it still takes 15 seconds to boot...
01:21:19rasherLlorean: that's just bizarre. I guess someone should hurry up and port j2me before the S port gets going anyway.
01:21:26stripwaxJdGordon - hm, that is strange indeed. Let me recreate the patch, that is not the version I actually have..
01:21:30rasherTraveler2: It says v2 on the back of it, and "audible support" on the box
01:21:42JdGordonlinuxstb: 15s from a hard disk player? thats not bad at all...
01:21:53amiconnJdGordon: That's sloow
01:21:55preglowrasher: sure, it does java natively. why not, it bloody has a simd fpu too
01:21:56Lloreanrasher: It claims to rival the effective throughput of a 3ghz system with its "Smart Speed Switch" technology. They seem to think this thing will solve all the world's problems too.
01:22:07amiconnArchos recorder boot within 5 seconds (with rockbox in flash)
01:22:16amiconnAnd that chip only has 11MHz
01:22:25JdGordonok, so maybe its only the h300 which is slow?
01:22:27*JdGordon shuts up
01:22:45amiconnThe H300 boot time is annoying to say the least
01:22:50n1sh300 OF takes between 30-60 seconds iirc :/
01:23:00Traveler2okay, so if has a sansa e200 and one of its images shows it running rockbox its a v1?
01:23:02linuxstbThe S's OF is almost 13MB though...
01:23:07stripwaxh120 OF takes longer, because of it's appalling tag database
01:23:14n1sthat was the major selling point for rockbox when I first got it :)
01:23:15zajacattackis there any work being done on an actual Rockbox player?
01:23:23LloreanTraveler2: As long as it's not the model that supports Audible.
01:23:42PaulJamh300 OF is only slow when the iriver db is enabled, for me it booted in ~12 seconds.
01:23:43n1sstripwax: the h300 takes about 30 with the db disabled...
01:23:47rasherSeriously though, who do I put in CREDITS for the matrix plugin? The guy who wrote the original cmatrix (at the bottom, perhaps)? The guy who ported it to ipl? The guy who fixed my mistakes (yes)
01:23:54amiconnJdGordon: _maybe_ an svn bootloader would speed things up a bit on H300. But I won't test one without a bdm wiggler around...
01:24:06Lloreanrasher: "Yes to all" I think?
01:24:06linuxstbrasher: "The Matrix team" ? (half-serious...)
01:24:06preglowrasher: well, was porting it to ipl major bother?
01:24:29amiconnPaulJam: Imo boot time on H300 *into rockbox* leaves a lot to be desired
01:24:30rasherpreglow: I've never looked at the original code, so I've no idea to be honest.
01:24:42JdGordonamiconn: I tihnk i have a testing bootloader on it... it doesnt really worry me seen as i dont use it anymore (its my brothers dap now)
01:24:44preglowrasher: if any part seems like a decent amount of work, credit the person who did it
01:24:47rasherlinuxstb: I'm leaning towards that, with a regular credit to the guy who fixed my code
01:24:49 Join Febs [0] (
01:25:04preglowthere is not "matrix team", so that would be silly
01:25:04rasherpreglow: He's not named by full name though :|
01:25:21rasherOnly "Alastair S" - maybe his name is known?
01:25:22PaulJamamiconn: well, the h300 is my only rockbox target, so i can only compare to the OF and tehre is see an improvement.
01:25:22JdGordonjust call him Bob
01:25:39rasherI'll just add everyone to CREDITS individually then
01:25:48preglowrasher: did he copyright the ipl source?
01:26:06rasherpreglow: There's a copyright notice with his name in it, below the original author's
01:26:13preglowok, then skipping his name is a bit too much
01:26:39stripwaxJdGordon - I'm an idiot, I reuploaded the old patch rather than the new one. Thanks!
01:26:49 Quit n1s ()
01:27:10rasherpreglow: I guess I'll just put Alastair S. It's not like he's a direct contributor, and I'll bet there are semi-anonymous contributors to some of our codecs as well
01:27:12JdGordonill commit it unless anyone has objections? (the real new version that is)
01:27:46JdGordon2 builds in a row crashed?
01:27:50Traveler2okay, has a sansa e280, any idea which version it would be?
01:28:02 Part toffe82
01:28:25preglowwhat, no one cared to mention the /var problem to the swedes?
01:29:08JdGordonbah bum!
01:29:11preglow/var has 150 megs of space left, i dunno if that's enough
01:29:16JdGordonat least it compiled...
01:29:41XavierGramiconn: boot time could be made quite fast on H300 by flashing the rockbox bootloader just like on H120/140, the thing is that we need a "Slahseri" (or should I say "Flasheri") with an H300 :P
01:29:52lostlogicNico_P: yay.
01:30:21preglowBagder: /var seems to be really tight on the build server
01:30:21lostlogicNico_P: you want to look into fixing gapless or want me to?
01:30:35Nico_Plostlogic: I won't do it now, that's for sure
01:30:46Nico_Plostlogic: do you have ideas?
01:31:41JdGordonstripwax: search_files() hits the hard disk right? wouldnt it be better to have albumart_get_filename() or something which returns the known filename instead of doinf the search again?
01:31:48JdGordonor wont it really make a difference for viewer?
01:33:12stripwaxJdGordon - it should only get called when you change the picture (e.g. select+right on ipod), it doesn't hit the disk more than necessary. also the "known filename" isn't quite right, I'm adding a note to the tracker to explain
01:33:17JdGordonstripwax: and last thing... you added functions into the middle of the plugin API which is bad!
01:33:31aliasklinuxstb: Did you ever get any of those random freezes that Nico_P had?
01:33:43JdGordonadd them to the end... or bump the plugin api version
01:33:59linuxstbaliask: No, I haven't got as far as trying to make the bootloader do anything interesting.
01:34:00stripwaxJdGordon- ah, crap. Thanks, yes, I'll add to the end.
01:34:04Nico_Paliask: I haven't seen them yet with linuxstb's patch
01:35:17aliasklinuxstb: Well, it seems they stemmed from the original nk.bin file that was being used. The ones on the wiki seem to cause that problem, but the one I've been using doesn't have the same issues
01:36:23 Join gromit`` [0] (
01:36:50 Quit gromit` (Read error: 110 (Connection timed out))
01:38:14Nico_Paliask: I've used the one from the wiki for dual boot, and linuxstb uses the same
01:38:23linuxstbaliask: BTW, is there a nicer place to get the OF nk.bin from? We probably shouldn't host them on the wiki....
01:38:54JdGordondoes anyone else think adding a new hour tag for the WPS to copy the config setting is a good idea? I hate to manually edit themes to fix it
01:39:08aliaskNot that I really know of. I would host them, but my IP changes way too often to be of any use
01:39:38stripwaxJdGordon - actually, why is adding a new fn into the middle of the api struct still a bad thing (do people really not just unzip the whole new build over the top, new plugins 'n all?)
01:39:39aliaskRouter doesn't support auto update
01:39:52aliaskAlso the DNS caching doesn't help
01:40:04JdGordonstripwax: yes
01:40:14stripwax:) ok then
01:40:19lostlogicNico_P: yeah, I think I can add a callback that's called whenever a handle is finished and a function that reclaims X bytes on the end of a handle
01:40:31 Join atsea-104 [0] (i=atsea-@gateway/tor/x-f3d05044e08097fa)
01:40:38stripwaxJdGordon - given that rbutilqt doesn't give you any other option, I assumed that was now ok
01:40:57pixelmastripwax: what's the fs number again?
01:40:58JdGordonnot everyone uses rbutil
01:41:04Nico_Plostlogic: I was vaguely thinking the same
01:41:21aliasklinuxstb: With my nk.bin the dual boot patch doesn't load anything, just sits at the first splash screen (before the windows mobile one)
01:41:25stripwaxJdGordon - true
01:41:31stripwaxpixelma -
01:41:32 Quit lucas42 (Remote closed the connection)
01:41:38stripwaxfor sliding puzzle that is (not pacbox)
01:41:40JdGordonstripwax: if you bump the api then its less bad, because at least then we wont have crashes when people use old .rocks witth a new binary
01:41:41zajacattackso, is there any way to run more than one rockbox build?
01:41:54stripwaxJdGordon - happy to move to end of api anyway
01:41:58*amiconn 's build server is on a dynamic IP as well, which changes at least once per day
01:42:02pixelmastripwax: thanks, does it always use the albumart?
01:42:43linuxstbaliask: Where did you get your nk.bin from?
01:42:55stripwaxpixelma - no. by default it uses a crafty Rockbox puzzle. Unless launched as a viewer, in which case it uses the file you select. You can cycle picture mode (default picture -> albumart -> numerals)
01:43:00linuxstbaliask: Or put another way, where can I get it from?
01:43:02aliasklinuxstb: I can't remember to be honest. I thought it was from the wiki, but obviously not...
01:43:22pixelmastripwax: ah, ok. Sounds neat.
01:43:40stripwaxer, groovy
01:43:40amiconnstripwax: Just take care that it'll still work on targets which don't have album art
01:43:52lostlogicNico_P: might need an additional read command "char *buf_get_location(int handle, ssize_t where, size_t bytes)" or some such where "where" can be positive for an offset from the beginning and negative for an offset from the end of a file
01:43:53stripwaxamiconn - it ought to
01:43:57lostlogicNico_P: or something like that
01:44:47JdGordonNico_P: is it possible to change the token which is used for a token string on load?
01:45:11Nico_Plostlogic: why not calculate that from the pointer you get with bufgetdata?
01:45:19stripwaxamiconn - although I am making the gross assumption that LCD_DEPTH==1 implies no albumart support in the build. I don't know if that assumption has changed now that albumart is in svn...
01:45:29Nico_PJdGordon: I don't see what you mean
01:45:40stripwaxI did try an archos build and an H120 build previously, but not recently..
01:45:42lostlogicNico_P: because bufgetdata may be operating at whoknowswhatoffset if the handle is currently being played by another client −− this is an out-of-band operation
01:45:43pixelmastripwax: I just don't want it to use my 40x40 album art I prepared as an experiment for my c200 :)
01:45:59 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
01:46:11linuxstbstripwax: I'm not sure but there may be a HAVE_ALBUM_ART define
01:46:15stripwaxpixelma - not sure I follow. I don't know the c200
01:46:29JdGordonNico_P: I want to use a different token value for the string depending on a global_setting value?
01:46:33stripwaxlinuxstb - hm.
01:46:55jhMikeSwhat's this gapless problem? I haven't had any trouble with it.
01:47:01amiconnlinuxstb might want album art on his logikdax
01:47:06*linuxstb checks and finds HAVE_ALBUMART
01:47:09stripwaxlinuxstb - is there anything intrinsic in albumart that would make you want to disable it in the build for color lcd devices?
01:47:19Nico_Plostlogic: good point
01:47:38 Quit Skail ("Concoction recent by Pier.")
01:47:43JdGordonNico_P: I want to get the wps to show 12/24 hour time depending on the config
01:47:46linuxstbstripwax: I don't think so.
01:47:52preglowlinuxstb: you know if that thing has a hardware mpeg decoder, btw?
01:47:55amiconn...with the greyscale lib built into the core ;)
01:48:04JdGordonand i thought the best way was to instert the right token when the .wps is being parsed
01:48:10stripwaxlinuxstb - should I predicate the calls to albumart_search_files on HAVE_ALBUMART?
01:48:11pixelmastripwax: just saying that it would be even smaller than what's currently used (c200's screen is only 132x80 pixels)
01:48:16linuxstbamiconn: I don't think the 128x64 screen could take it... ;)
01:48:22JdGordon.. but that wont change at runtime if the setting changes, so i tinhk ill do it a diferent way
01:48:28 Quit Febs (Remote closed the connection)
01:48:54Nico_PJdGordon: it's not a string setting
01:49:10stripwaxpixelma - the plugin takes the bitmap and resizes it to fill the screen - it wouldn't use the "wrong" resolution in any way, if that's what you mean. But if your only albumart is 40x40 then, yes, it will use your 40x40 bmp and resize it (not clear what else it could do in that situation..)
01:49:11pixelmastripwax: and so I experimented with really small bmps
01:49:29Nico_PJdGordon: unless I didn't understand correctly (it's late), I think you want a new token
01:49:45stripwaxpixelma - my note at the bottom of the tracker page should be useful?
01:49:45JdGordonyeah, but i dont want to add a new token
01:49:47XavierGrsh*t! power outage!!
01:49:48linuxstbpreglow: Which thing?
01:49:50JdGordonbut i tinhk ill have to
01:49:53 Quit XavierGr ("One firmware to rule them all!")
01:50:12jhMikeSbuffering need serious simplification
01:50:18Lloreanpreglow: It claims a hardware encoder, at least
01:50:19Nico_PJdGordon: do you want to use the same tags/tokens as the existing RTC ones?
01:50:30preglowlinuxstb: that logik dab thing
01:50:34Nico_PjhMikeS: how?
01:50:57JdGordonNico_P: dont worry.. its not as bad a hack as i was scared of
01:50:58 Join webguest57 [0] (i=63e88a81@gateway/web/cgi-irc/
01:51:28jhMikeSNico_P: do you allow memory reservation of greater than ram size?
01:51:41webguest57Is there an easy way to search a data file for a string in rockbox?
01:51:41Lloreanpreglow: "MPEG4 Encoder", I'm assuming it means video.
01:51:41 Quit PaulJam (".")
01:51:49pixelmastripwax: I'm fine if I can just let it not use those bmps (so either using the "built-in" one or another bmp)
01:52:08*jhMikeS finds the code a bit difficult to follow in any case
01:52:10linuxstbpreglow: No, just the Telechips SoC. Although I'm not 100% sure what's inside the DAB module, although I wouldn't expect a hardware decoder.
01:52:35preglowlinuxstb: no, me neither
01:52:48linuxstbaliask: Any idea how to power-off the S in software?
01:53:13preglowlinuxstb: docs on the dab module?
01:53:43 Quit Traveler2 ("Java user signed off")
01:54:12stripwaxpixelma- yep, and that's the default behaviour. but having said that I haven't tried building for c200, it might just look plain ugly.
01:54:56Nico_PjhMikeS: a file that is bigger than the buffer will prevent another one from loading, so I'd say yes
01:56:00jhMikeSwhat I meant was reserving 200MB or something for a track explicitly making buffer space logically infinite
01:56:29Nico_PI'm not folowing
01:56:37amiconnrasher: loadsa red...
01:56:43 Quit webguest57 ("CGI:IRC (EOF)")
01:57:11JdGordonNico_P: Llorean: others: any objections to ?
01:57:58linuxstbaliask: I've just tried with your nk_orig.bin, and the dual-boot worked fine.
01:58:11jhMikeSto simply reserve x number of bytes in advance on the buffer as a placeholder for data and in no way limiting this allocation to RAM size.
01:58:21aliasklinuxstb: Power off can't really be done - but you can put it to sleep. IIUC both the battery, and the power cable are tied to interrupts which will turn the unit on, so as long as the battery switch is on, it can't turn off
01:58:33linuxstbaliask: How annoying...
01:58:33aliasklinuxstb: Odd. Maybe I did something wrong...
01:59:47aliasklinuxstb: I patched mknkboot.c and, rebuilt the bootloader.bin, rebuilt mknkboot, used mknkboot on my nk_orig.bin, then sent the resulting nk.bin to the player. Did I miss anything?
01:59:49Nico_Paliask: does it display the progressbar?
01:59:57aliaskNico_P: Nope
02:00:00 Join XavierGr [0] (
02:00:24jhMikeSanother is using data attributes instead of types...say DATA_HAS_POINTERS, DATA_ATOMIC, etc. in order to not special-case for types
02:00:25Nico_Paliask: that would happen to me with a totally bogus nk.bin
02:00:46LloreanJdGordon: Honestly, I'd be in favour of getting rid of the old hour tags too. But that's me, and we probably shouldn't break WPSes yet again. :)
02:00:52LloreanBut yeah, no objection
02:01:08JdGordonjust checking the code actually works :p
02:01:25LloreanGood plan
02:01:55linuxstbaliask: Maybe I've changed something in my dual-boot patch, let me upload another.
02:02:51Nico_PjhMikeS: there is no real reservation. when trying to add something to the buffer (bufopen of bufalloc), we simply check if there is enough space left to finish buffering the current file. I'm not sure what you're saying, are you suggesting adding a reservation mechanism?
02:02:56JdGordoncrap.. stupid sim is crashing
02:03:12aliasklinuxstb: Got it working. I think I gave mknkboot the arguments in the wrong order before
02:03:16 Quit scorche (Nick collision from services.)
02:03:32linuxstbaliask: Ah, that would do it...
02:03:45 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
02:04:00stripwaxJdGordon - ok, patch updated, but I haven't had a chance to try building for all possible targets..
02:04:32JdGordonyou done a target with no AA?
02:05:07jhMikeSNico_P: sort of I guess. meaning space for the whole file would be reserved in advance.
02:05:25JdGordonlame.... you cant change the clock in the sim...
02:05:35stripwaxJdGordon - not yet, but it's pretty late here. If nobody else will have a chance to test, I can take a look in the next day or so
02:05:36JdGordon12pm is a really sucky time to be testing this patch!
02:05:45stripwax1am is worse :)
02:05:54JdGordonno... 1am is 13 in 24 hour time
02:05:55*preglow blinks at his 2am
02:06:00JdGordon12 is always 12
02:06:17jhMikeSJdGordon: around -2pm or so ?
02:06:36LloreanJdGordon: 1am is 1 in 24-hour... it's 1pm that's 13...
02:06:51safetydanpreglow: no, haven't got further than the initial discussion. Now that the hardware eq is out though, I might start looking at it again.
02:06:55stripwax12 is sometimes 0
02:06:57*JdGordon needs more sleep
02:07:02*stripwax does too
02:07:04stripwaxnight all
02:08:07linuxstbaliask: I've just had a look at the tools/gigabeats.c code - I assume that simply creating two records (the eboot patch, plus the bootloader) doesn't work?
02:08:08jhMikeS0 = 12 midnight, 1 to 11 = 1 to 11 am, 12 = 12 noon, 13 to 23 = 1 to 11 pm :P
02:08:36*preglow implodes
02:08:54aliasklinuxstb: Well, it may now that we send the nk.bin file directly rather than using the gigabeat v updater
02:09:04aliaskBut before it was rejecting it before it even got to the player
02:09:41aliaskWhich is why that gigabeats.c file has so many strange entries, I was trying to get it to work by copying chunks from the OF nk.bin
02:10:04amiconnrasher: lazy, eh? ;)
02:10:25rasheramiconn: absolutely. How did I get caught?
02:10:33linuxstbNico_P: What's the status of your experiments with identifying files transferred via libmtp?
02:10:34 Join iamben_ [0] (
02:10:41rasheramiconn: Oh, not converting to grayscale?
02:10:48 Nick fxb is now known as fxb__ (
02:11:01rasherI wasn't planning on doing that, really.
02:11:19rasherI doubt it should be hard. There's really only a single drawing call
02:11:27linuxstbrasher: Is any conversion needed?
02:11:39*jhMikeS is about to inflict real damage on things ;)
02:11:55Nico_Plinuxstb: I've found them, but the names are seq numbers
02:12:22linuxstbThat's to be expected...
02:12:25rasherlinuxstb: no idea. I really don't know how the bitmaps system works to be honest.
02:12:40 Join ChanServ [0] (ChanServ@services.)
02:12:40Mode"#rockbox +o ChanServ " by
02:12:51amiconnrasher: Looks like that for greyscale, all that's missing is a suitable bitmap. Plus keypad definitions for H1x0 and ipod G1..G3
02:13:17Nico_Plinuxstb: but I think we can reasonably assume that the first file we send will be 02
02:13:56linuxstbIs Rockbox itself building for the S?
02:13:59rasheramiconn: ah, the bitmap isn't converted to greyscale if used on a greyscale target?
02:14:05Nico_Phaven't tried :p
02:14:20*linuxstb sets a build going now...
02:14:21amiconnrasher: The bitmap isn't compiled atm, that's why liniking fails
02:14:49*aliask races linuxstb
02:14:58Nico_PjhMikeS: I'd be interested in discussing your idea, but tomorrow... I really need to go to bed now
02:15:19linuxstbaliask: It almost builds - some functions in the debug menu don't exist...
02:15:21Nico_Plinuxstb, aliask: I can give you my recursive dir listing if you want
02:15:21rasheramiconn: ah, yes. And the button problem.
02:15:21jhMikeScommitting this kernel stuff always feels like skydiving :\
02:15:38Nico_Pdo you skydive? :p
02:15:39jhMikeSNico_P: gnight
02:15:47amiconnThe button "problem" is dead simple to solve
02:15:49linuxstbaliask: But the linking address (0x80000000) in firmware/ is wrong - IIUC DRAM starts at 0x88000000
02:15:56jhMikeSI should...less stressful :)
02:15:56Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
02:16:14aliask0x8000000 is ROM I think
02:16:28 Quit hcs ("Leaving.")
02:16:57aliaskI must have just copied the datasheet when I was writing that.
02:17:24amiconnJust use the IPOD_4G_PAD definitions for IPOD_1G2G_PAD and IPOD_3G_PAD, and the IRIVER_H300_PAD definitions for IRIVER_H100_PAD
02:17:29linuxstbaliask: My dualboot patch now loads the bootloader to 0x8a000000, so plenty of room to load Rockbox at 0x88000000
02:17:38Mode"#rockbox +eeee *!*@rockbox/administrator/scorche *!*@rockbox/developer/amiconn *!*@rockbox/developer/lostlogic *!*@rockbox/developer/preglow " by scorche (n=scorche@rockbox/administrator/scorche)
02:18:22amiconnWe distinguish these keypads because of the different physical button arrangements, but apart from one single plugin, there are no actually different assigments so far
02:18:26preglowscorche: point being?
02:18:42 Join Febs [0] (
02:18:43scorcheexceptions in case of attack...handy to have at times
02:18:50rasheramiconn: already done that, in fact
02:18:54rasherTesting now.
02:19:00JdGordonscorche: why not except all rockbox/developer/* ?
02:19:20amiconn(the single plugin is the jpeg viewer, where the buttons for flipping through the folder are swapped between H300 and H1x0)
02:19:20Nico_PJdGordon: I might be late but I have no objection for your patch
02:19:28scorcheexceptions only really make sense when the people being excepted can do something about said attack
02:19:39JdGordonNico_P: great :), just commited
02:19:40preglowscorche: if we're ever attacked, that'll be pretty easy for any attacker to reverse, no?
02:20:30*jhMikeS pushed the big red button
02:20:32*preglow doesn't know all the freenode bells and whistles
02:20:33scorchemost people dont know about it, and it isnt scripted commonly like kb is...either way, it gives time to respond...there really is no downside to it, and after they are set, you wont notice anything
02:20:47Nico_Plinuxstb, aliask:
02:20:59Nico_Pit includes the dual boot patch
02:21:03Mode"#rockbox +eeee *!*@rockbox/developer/T0mas *!*@rockbox/developer/Bagder *!*@rockbox/developer/Zagor *!*@rockbox/developer/LinusN " by scorche (n=scorche@rockbox/administrator/scorche)
02:21:24aliaskNico_P: Nice
02:21:28Mode"#rockbox +e *!*@rockbox/administrator/Febs " by scorche (n=scorche@rockbox/administrator/scorche)
02:21:37Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
02:21:57linuxstbaliask, Nico_P: Have either of you tried writing to the disk?
02:22:05*Nico_P hasn't
02:22:11aliasklinuxstb: I did once, and it didn't work
02:23:27linuxstbaliask, Nico_P: Should I commit this dual-boot stuff?
02:23:45 Quit iamben (Read error: 110 (Connection timed out))
02:23:52aliasklinuxstb: I say yes
02:23:56jhMikeSlinuxstb: call it "dual-day"
02:24:09aliaskNico_P: Can you apply that git patch using regular patch?
02:25:25rasheramiconn: well that works just dandy. Can't check it on target though. Sold my h120
02:25:30*linuxstb commits and goes to sleep
02:25:34*jhMikeS has to wait through a build blocking him :\
02:25:43*JdGordon appologises...
02:25:43scorchelinuxstb: dangerous practice :P
02:26:08rasherscorche: but this commit's bugfree.
02:26:18Nico_Paliask: I think so
02:26:19jhMikeSthey all start out that way :)
02:26:20JdGordonarnt they all?
02:26:39Nico_Paliask: use -p1
02:26:50aliaskAh, of course...
02:28:58linuxstbJdGordon: I guess I'm too slow, but how about a WPS tag that says what the 12h/24h setting currently is? So the WPS could display "22:00" or "10.00pm" depending on that setting.
02:29:45JdGordonhave we got an am/pm tag already?
02:29:52*jhMikeS sees the motley combinations of commits in the current build :p
02:29:58JdGordonyeah, thats a good idea
02:30:14*preglow watches free space on /var grow smaller and smaller :/
02:30:31jhMikeS /var ?
02:30:41rasherAt least we squeezed 3 commits into a single build.
02:30:42preglowthe build server /var partition is almost full
02:30:57preglowit currently has 130 meg left
02:31:00preglowjhMikeS: the main one
02:31:12JdGordonshe'll be right...
02:31:12linuxstbJdGordon: And then you wouldn't need %cf or %cg...
02:31:14jhMikeSout of what?
02:31:29*Nico_P is really off to bed now
02:31:29preglowjhMikeS: 19 gigs
02:31:32 Quit Nico_P (Remote closed the connection)
02:31:38*linuxstb is really off to sleep too
02:31:39JdGordonlinuxstb: but haveing the one tag instead of the conditional is easier to write up
02:31:40jhMikeSonly 19 gigs?
02:31:55preglowjhMikeS: looks like it
02:32:10amiconnjhMikeS: Is there a special reason for the moved .stack in the .lds?
02:32:15preglowand there it crashed into 0kb
02:32:59jhMikeSamiconn: put it at the end? any reason the smaller stacks should go after the main one?
02:33:25amiconnNo. Just asking as that seems to be not really necessary
02:34:50zajacattackout of curiosity, why is this project compiled with linux and yet designed for FAT32?
02:35:26krazykitzajacattack, because fat32 is supported on nearly every platform out there
02:35:46amiconnThere's tons of additional reasons too
02:36:00krazykitnot only that, for players that dual-boot, you can't format it something that the OF can't read: this is especially true on iPods and Sansas, where you need the OF to transfer files
02:36:27amiconnAnd btw, rockbox can be compiled on a variety of OS
02:38:36jhMikeSamiconn: It's partially that I'd rather have those stacks not be subject to damage from overflows on the main one if that occurs.
02:38:57amiconnSeveral targets couldn't even boot if the disk wouldn't be fat formatted
02:38:58jhMikeSactually, that's is the reason
02:39:44preglowlooks like that went well
02:39:52preglowapart from the no space left builds
02:39:58amiconn(from head: all archoses when not flashed, the iaudios, and the .mi4 targets at least)
02:40:32amiconnrasher: red again...
02:40:50amiconnmono != greyscale
02:40:59preglowhm, why is the 1-3g binaries so much smaller than the rest of the ipods? bitmaps?
02:41:17rasheramiconn: I know. I even put that in the bitmaps SOURCES... I'm just on crack, it seems.
02:41:29rasheramiconn: Put LCD_DEPTH > 1, that is
02:42:03alienbiker99does the new gigabeat s bootloader make you erase your hd?
02:42:28jhMikeSactally, I did exactly the opposite of what I wanted to do. :\
02:43:05amiconnpreglow: Bitmaps and features. The G3 and G4 should be equal bitmap wise. But G4 has usb stack and recording. G3 has neither of those
02:44:29preglowbut bedtime for me
02:45:08amiconnjhMikeS: You mean the stack?
02:45:39jhMikeSyeah, I'm thinking backwards here.
02:46:23amiconnDoes dualcore work nicely now? I guess you tried spc?
02:46:35 Join DM| [0] (
02:46:41 Join LycoLoco [0] (
02:46:43jhMikeSyes, listened to it for hours
02:47:12amiconnmpegplayer should also use the second core?
02:47:17jhMikeSit does
02:47:41amiconnOk, will try that out when my new greyscale stuff is taking shape :)
02:47:42jhMikeSthe graylib interrupt can still cause some audio skipping (short though)
02:48:08 Quit Thundercloud (Remote closed the connection)
02:48:14amiconnBut now, sleep
02:48:23jhMikeSI guess if graylib runs faster it won't though
02:50:55 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
02:51:36rasherWhat, more red?
02:51:51rasherAh, not my commit. More waiting, then.
02:57:29jhMikeSall matrix stuff
02:57:40 Join lucas42 [0] (
02:59:43jhMikeSpreglow: ipl has an SPC emulator? not as up to snuff as ours you said?
03:04:37pixelmawasn't it hcs who mentioned it /me thinks in half asleep state
03:05:25 Part pixelma
03:05:40 Quit TTThomas ("Lost terminal")
03:09:33 Part zajacattack
03:10:56 Join darkapostrophe [0] (
03:13:44***Saving seen data "./dancer.seen"
03:23:07 Join psycho_maniac [0] (
03:23:30 Join Phoul [0] (i=AOS@sourcemage/guru/Phoul)
03:23:49PhoulHas anyone successfully installed rockbox onto a 3rd gen ipod nano video
03:24:01 Join dvdv [0] (n=ut@
03:24:21psycho_maniacif you look on the main page Phoul you will see if your player is supported by rockbox.
03:24:32PhoulI know its -not- supported lol
03:24:47PhoulBut just because something isnt "supported" doesnt mean theres no way of getting it to work ;)
03:24:58psycho_maniacthen the answer to your question is probably "no"
03:25:04psycho_maniaccheck the wiki and the new port topics
03:25:22 Part Phoul
03:26:36rasherActually, it lists targets on which Rockbox runs. You would think that's specific enough.
03:27:43 Quit Dark_Apostrophe (Connection timed out)
03:27:44 Nick darkapostrophe is now known as Dark_Apostrophe (
03:27:45Lloreanrasher: People seem to think it's a case of simply "try and see if it works"
03:27:50psycho_maniacok this is strange. on my ipod 5.5g the backlight will not turn on
03:28:01Lloreanrasher: Like the people who ask what file from their favorite N64 emulator they should rename to .rock to get it to run.
03:28:17Lloreanpsycho_maniac: What do you have the brightness set at?
03:28:26rasherLlorean: Oh dear. Are you serious?
03:28:43Lloreanrasher: 100%.
03:29:05*jhMikeS shakes head
03:29:05Lloreanpsycho_maniac: I don't know if that's bright or dark, but I just meant 'maybe it's dim enough that it looks off'
03:29:20psycho_maniacupdated both my gigabeat and ipod 5.5g at the same time and this only happens on the ipod
03:29:24psycho_maniac16 is the default
03:29:31psycho_maniacon the ipod
03:29:36jhMikeSreset settings?
03:30:36psycho_maniacthat worked. now i just have to find out what makes it do that.
03:34:06psycho_maniacahh who cares. i dont like the ipod anyways.
03:38:42 Part dvdv
03:46:04 Quit japc (Read error: 104 (Connection reset by peer))
04:07:44 Quit donutman25 ("ChatZilla 0.9.79 [Firefox]")
04:13:09 Join sdoyon [0] (
04:15:37 Quit advcomp2019 (Read error: 110 (Connection timed out)) question: when make voice is run a second time, why not overwrite english.voice? Why are we creating an english-1.voice? Took me a while to notice and figure out why my voice file was wrong :-).
04:21:30 Join stripwax [0] (
04:24:01 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
04:25:04 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
04:26:23 Quit advcomp2019 (Nick collision from services.)
04:26:33 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
04:30:49sdoyonI get an SQL error with the tracker!
04:31:20 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
04:32:01 Join linuxstb__ [0] (
04:32:21aliasksdoyon: The tracker is working for me, what page are you trying to access?
04:33:07aliaskOh wow! Yes, that's certainly broken.
04:33:16sdoyonFailed! (Incorrect key file for table '#sql_3edd_0.MYI'; try to repair it)
04:33:56aliaskBagder: Ping
04:34:07sdoyonWell thanks for letting me know I'm not crazy :-)
04:35:14 Join Simon14 [0] (
04:35:16sdoyonWould you happen to know the URL to add a new task?
04:36:27 Join iamben [0] (
04:36:51 Join TTThomas [0] (
04:38:08sdoyonaliask: hmm thx, but that's not working either for me. Thought I just saw a notification for a new open bug go by the mailing list. Oh well no matter.
04:38:39aliasksdoyon: That link works for me, but I had to log in first
04:40:28sdoyonaliask: strange.. Hmm OK, at what URL could I try to login?
04:40:50*jhMikeS questions whether that error info should even be displayed publicly :\
04:41:41aliasksdoyon: All tasks should have a login box at the top. For example you could use
04:41:59aliaskjhMikeS: I don't see what's sensitive about it...
04:42:03jhMikeSit dumps the whole query? ouch.
04:42:20 Join billenium [0] (
04:42:29jhMikeSI don't know either atm.
04:42:38aliaskWell, it's not elegant, but it's not going to compromise security.
04:43:28sdoyonaliask: That task works, another one I was trying did not for some reaspon... Perhas I had an invalid id...
04:43:39 Quit linuxstb (Read error: 113 (No route to host))
04:43:41jhMikeSbetter to never show it imo
04:44:07 Quit stripwax (Read error: 110 (Connection timed out))
04:45:21sdoyonaliask: Thanks, appears to work.
04:49:19 Quit linuxstb_ (Read error: 110 (Connection timed out))
04:50:20 Quit iamben_ (Read error: 110 (Connection timed out))
04:50:33 Quit DM| ("*bashes head against keyboard*")
04:57:29 Join keanu [0] (n=keanu@unaffiliated/keanu)
04:58:05keanuI appreciate being able to see the entire mysql query on the tracker, but really, I don't need to see it ;)
04:59:02 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:59:13 Join miepchen^schlaf [0] (
04:59:59 Join Norrin [0] (i=4bce7617@gateway/web/cgi-irc/
05:01:00NorrinGood evening all.
05:01:56 Join karashata [0] (n=karashat@
05:02:41 Quit Llorean ("Leaving.")
05:03:31keanuAnyone have access to fix mysql?
05:03:53NorrinI'm asking for some help concerning my Archos FM recorder I recently upgraded to the newest build.
05:04:54NorrinWhen I boot it straight up, it gives me an ATA error: -11. I can plug it in to my USB and it works fine as an external HD. I can then remove it and it boots normally. It just seem to have a problem with the cold boot and shutdown.
05:05:06NorrinAny help is greatly appreciated.
05:06:35 Quit J3TC- (".•«UPP»•.")
05:07:02keanuNorrin, hold on, i'm trying to figure out what error -11 is
05:07:45keanuNorrin, this might help until I find an answer:
05:09:29keanuNorrin, according to the mail archive, it means it's because the hard drive isn't accessible or doesn't spin up.
05:09:47 Quit karashata ("I will *SO* make you regret that... Later... *is a lazy dragon*")
05:10:32 Quit Norrin ("CGI:IRC (Ping timeout)")
05:11:17 Join Norrin [0] (i=4bce7617@gateway/web/cgi-irc/
05:11:25keanu"Norrin, this might help until I find an answer:"
05:11:27keanu"Norrin, according to the mail archive, it means it's because the hard drive isn't accessible or doesn't spin up."
05:11:30 Join homielowe_ [0] (
05:11:43keanuin case cgi:irc timed out before those two messages ;)
05:11:48NorrinSorry, got booted off for some reason.
05:11:57 Nick homielowe_ is now known as homielowe (
05:12:16keanunot a problem
05:13:24psycho_maniacis there any talk about removing the first screen that comes up when you turn on the gigabeat Fxx series?
05:13:47***Saving seen data "./dancer.seen"
05:15:29NorrinIt worked fine earlier today with a previous build. Any clue why it would suddenly develop a problem immediately after upgrading?
05:16:12keanuit could have been a failing drive I assume, though I believe you did say you can still access it as an external HD, right?
05:16:36NorrinIf I plug in to the USB first, let it boot in USB mode, then remove the USB, it works as advertised.
05:17:00psycho_maniacif you believe it is because of the build you are using i would try a older build. then if it still does it it would be for sure a hd failure like keanu said.
05:17:06 Quit jurrie ("Leaving")
05:17:33keanu*it *could* be for sure a hd failure
05:17:39NorrinI also can run the Archos 1.30j file without any errors.
05:18:18keanuIn that case, try an older rockbox build and see if it works, and maybe try reseating the battery.
05:18:24psycho_maniactry an older build to see if you get this error
05:19:09keanuI don't have the FM recorder (or a HDD-based player - I have a Sansa) so I'm just going off of what I'm seeing in the mail archive and forums
05:19:18 Quit Simon14 ("ChatZilla 0.9.79 [Firefox]")
05:19:34*keanu rants for a minute about flyspray still being down
05:19:36NorrinI'll try the 2.5 build and see what happens.
05:20:14 Join jurrie [0] (
05:22:33NorrinThanks. Loaded the 2.5 and it worked fine. Need to find which build causes problems. Like the newer capabilities (except for the ATA error.)
05:22:48NorrinThanks for your time.
05:23:17keanuGlad I could help. Once you find which build it is, you may want to report it as a bug
05:27:54 Quit Norrin ("CGI:IRC (Ping timeout)")
05:32:00 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
05:32:14 Quit advcomp2019 (Nick collision from services.)
05:32:20 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
05:35:16 Quit sdoyon ("ircII EPIC4-2.4 -- Are we there yet?")
05:35:50 Join Norrin [0] (i=4bce7617@gateway/web/cgi-irc/
05:38:03Norrin20071102 works, 20071109-06 have not. Trying 1104 now.
05:43:20 Quit Norrin ("CGI:IRC (Ping timeout)")
05:46:53 Join Norrin [0] (i=4bce7617@gateway/web/cgi-irc/
05:52:19 Quit Norrin ("CGI:IRC (Ping timeout)")
05:56:02 Join Ebert [0] (
06:07:30 Quit DaCapn (Read error: 110 (Connection timed out))
06:13:12 Join DaCapn [0] (
06:15:34 Join JdGordon_ [0] (
06:16:21 Quit JdGordon (Read error: 104 (Connection reset by peer))
06:18:22 Quit DaCapn ()
06:28:59 Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
06:43:55 Join l7 [0] (
06:44:32l7anyone have an opinion on which mp3 player is the best rockbox platform?
06:44:56psycho_maniacl7: have you checked the buyers guid?
06:44:57l7i'm trying to decide between sansa e200 or c200 series
06:45:07l7which buyer's guide?
06:45:18l7is it on
06:45:41Inkey5I just got the e200 today
06:45:42l7cnet seems to claim that the c200 series has poorer sound quality
06:45:48Inkey5Im pretty happy with it
06:46:00l7ah thanks for that link
06:46:15l7i also heard that more of the dev team has e200s
06:46:37Inkey5it being a bit newer and alot more popular, I'd expect that to be the case
06:46:48l7hrm yeah
06:47:06l7the bigger screen on the e200s seem to be a nice bonus
06:47:34psycho_maniacthere a pretty good price too arent they?
06:47:37l7sandisk advertises it's 1.8 inch screen but is mysterious about the c200's exact size
06:47:43safetydanbe aware that there's a version 2 of the e200 and c200 that Rockbox currently doesn't work.
06:47:45Inkey5I got mine off woot for $30
06:47:45l7yeah, they're both pretty cheap
06:47:49l7the c200 is a bit cheaper
06:47:51safetydanwork on even
06:48:05l7i missed out on that woot
06:48:12Inkey5only gotcha is no usb yet in rockbox
06:48:13psycho_maniacInkey5: how much space does it have when you bought it.
06:48:22Inkey5gotta boot into OF to load stuff
06:48:29Inkey5plus micro-sd slot
06:48:37l7USB support is coming eventually right?
06:48:42psycho_maniacwhich can hold up to 5gb? for how much?
06:48:54Inkey5"probably" is the answer I got
06:48:59l7i'm not sure if i read the c200 faq properly, but is there USB for the c200?
06:49:08Inkey5psycho_maniac: I suppose so; it was $30
06:49:17Inkey5I think so
06:49:26Inkey5I think the c's had a built in usb bridge
06:49:32psycho_maniaci take it you didnt get a msd card?
06:49:46l7maybe you could just transfer files from the microsd card if you wanted to avoid using USB someimes
06:50:19Inkey5ooooh, according to the faq the e200 supports micro-sd up to 32gb
06:50:43l7wow, the e200 has 15h of batter life compared to 5-6h for the c200
06:50:57Inkey522 if you stick with stock firmware
06:51:33l7only thing about the e200 is the lack of a charger
06:51:53Inkey5psycho_maniac: I have a 1gb from my phone, but havent bought a bigger one yet can buy those, though
06:52:04DogBoyI have a 2 gig in mine
06:52:17Inkey5and how often are you going to go 16 hours without being by your computer?
06:52:41DogBoyI've seen car chargers for them too
06:52:44Inkey5and even so, they have batteries you can swap in the field with only an eyeglass screwdriver
06:52:54Inkey5or a swiss army knife
06:53:10l7can you put sdhc cards in either sansa?
06:53:43advcomp2019but the OF will not read them
06:53:52Inkey5it read mine ok
06:54:01l7Inkey5: which one do you have?
06:54:05Inkey5even with my wierd motorola filesystem from my phone on it
06:54:21Inkey5also,I cant speak for the c200's, but the screen on the e250 looks really nice
06:54:32l7advcomp2019: living w/o OF support for SDHC shouldn't be much of a problem right?
06:55:02l7the only nit is i guess it means you will need to use a card reader until rockbox has usb support
06:55:09Inkey5what would be nice would be a plugin for RB to copy all media from the sd to the flash
06:55:22tuplanollamy friend has c200 and it's screen is just awful when compared to my e280
06:55:24Inkey5l7: dual-boot into the OF and do your transfering
06:55:51l7Inkey5: you have a e250, not an e250R right?
06:56:03advcomp2019l7, you just need figure out a way to transfer stuff to microsdhc card
06:56:10l7i guess i should just go and get one
06:56:17Inkey5its a good player, though
06:56:25Inkey5the picture viewer in RB is pretty sweet, too
06:56:29l7advcomp2019: it should be okay, as long as an exotic filesystem isn't needed
06:56:37Inkey5it shows my 1600x1200 pictures from my camera
06:57:02Inkey5scaled down, but you can zoom in all the way and pan around
06:57:07Inkey5its pretty nice
06:57:17l7just wait until the larger sdhc cards come out...
06:57:22l7so much for ipod
06:57:42Inkey5clearly youve not played with an ipod touch
06:58:10Inkey5its pretty bad ass
06:58:10l7actually i have
06:58:23l7i'm about to get a nokia n800 though :)
06:58:24Inkey5it could conceivably replace my laptop for a lot of things
06:58:53l7i think the n800 would be more useful than the touch for most things, and the e250 would be better for working out or whatever
06:59:11l7Inkey5: the ipod touch has no bluetooth though...
06:59:25l7well it might have a disabled bluetooth chip inside
06:59:34scorchethis is getting offtopic...
07:00:06Inkey5but you can jailbreak them and add any number of apps
07:00:07Inkey5like mail
07:00:11Inkey5or office
07:00:20safetydanway off topic
07:00:21scorchethis is offtopic...
07:00:33Inkey5is that bad?
07:00:43*scorche points to the guidelines in the topic
07:00:51l7well i guess we could yack in #rockbox-community
07:01:15*psycho_maniac was wondering when somebody was going to say something
07:01:24Inkey5sorry then, chaps
07:01:31Inkey5well in short, get an e200
07:02:14 Quit BHSPitMonkey ("Leaving")
07:06:36 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
07:08:26 Join ddalton [0] (
07:08:55 Quit ddalton (Client Quit)
07:09:30 Join ddalton [0] (
07:13:51***Saving seen data "./dancer.seen"
07:13:58 Join perrikwp [0] (i=98216687@gateway/web/cgi-irc/
07:19:26 Quit perrikwp ("CGI:IRC (Ping timeout)")
07:23:47 Quit linuxstb__ (Read error: 104 (Connection reset by peer))
07:23:54 Join linuxstb__ [0] (
07:28:22l7does anyone know if the microsd card for e200s causes a loading delay in rockbox?
07:28:47l7i have an older sansa and the loading delay is very annoying
07:29:12Inkey5only one time as it refreshes the database
07:30:25advcomp2019l7, i tried a a friends card in my e280r with rockbox and i had no delay, but i am getting my own card to play around with
07:31:50 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
07:31:53Inkey5it doesnt slow it down just being in there, though, if thats what you mean
07:32:17 Join miepchen^schlaf [0] (
07:33:54l7Inkey5: the annoying thing about the older sandisks is that if you used them with a SD card, it would take about 2 minutes to load songs from the SD every time you booted it up
07:34:28l7the longest it could standby for was about an 1/2 hour, which was troublesome
07:34:31psycho_maniacis this with the OF or with rockbox?
07:34:40l7psycho_maniac: with the OF
07:34:52l7this is a sansa e130, really old :)
07:35:05Inkey5nah, not a problem
07:35:11Inkey5it only refreshes the database on changes
07:35:16l7how long does the e200 series take to boot up from being off?
07:35:18l7that is smart
07:35:21Inkey510 sec even
07:35:29Inkey5to OF
07:35:31 Quit aliask ("ChatZilla 0.9.79 [Firefox]")
07:35:34Inkey5rockbox is a bit faster
07:36:07 Quit ddalton ("leaving")
07:36:13l7yeah my creative zen takes about 10 seconds as well
07:36:31l7how long can the e200 stay in standby mode before it powers off?
07:36:34l7in rockbox
07:36:40Inkey5I think thats a configurable setting
07:36:56psycho_maniachow is the sound in the h120 compared to the gigabeat Fxx?
07:37:02Inkey5up to 60min
07:37:05l7i wonder how much battery life it would have idling and not playing songs
07:37:57l7one of the things i really liked about my ipod before it died was that it would start up really fast from sleep mode
07:38:02psycho_maniaci always hear about people having problems with there ipods when the battery power gets really low. would i risk this happening if i ran a the battery bench plugin?
07:39:15DogBoyof course
07:39:56scorchepsycho_maniac: just charge it afterwards
07:41:42l7hrm, according to, "Not all plugins have been adapted to the Sansa's screen layout yet."
07:41:55l7any idea what plugins these are?
07:42:19l7i could probably live w/o doom at optimal res :)
07:43:10psycho_maniacif nobody gives you an answer the easiest way is to go to the plugins wiki page and see what plugins are missing on your player compared to that wiki page.
07:43:56Inkey5actually doom works great
07:45:03 Quit Inkey5 (Read error: 104 (Connection reset by peer))
07:45:40l7i guess i can live with whatever the e200 is missing
07:45:42 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
07:46:17l7the wiki also mentions that you could flash the e200 with e200r firmware... has anyone tried this one?
07:47:20scorchewhy would you want to?
07:47:24l7how does it work?
07:47:46l7i thought it might be neat to be able to use rhapsody on the go
07:47:54 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
07:48:12l7i would probably get bored of it and go back to rockbox though, but it would be a nice option
07:54:29 Join OlivierBorowski_ [0] (
07:58:03scorcheLinusN: want to nuke this person's avatar?;u=13125
07:58:55LinusNbecause it's a screenshot?
07:59:11scorche"nigga stole my yoshi"
07:59:26LinusNah, now i see it
08:00:02LinusNhmmm, tough call
08:01:01scorcheif we discourage and censor the use of "spaz" which is offensive in the UK, but not the US, i dont see how "nigga" is any better
08:01:22scorchealthough, others are free to input..
08:01:26LinusNwe do?
08:01:48LinusNlet's pm him and ask him to change it
08:01:49hcsdoes anyone consider "nigga" offensive?
08:02:06LinusNi dunno, you tell me
08:02:23scorchehcs: well, considering its direct root..
08:02:24hcsthey tell me I can't have an opinion if I'm not a minority
08:02:41hcsit is in very common usage in an inoffensive sense, though
08:02:58hcsthough here we have it specifically applied to a black man
08:03:03hcsby a white man
08:03:13hcs(well, sprites)
08:04:08scorchealthough i suppose we can always wait for Llorean's word on it..
08:04:29scorchei would just think it would be something considered offensive
08:06:05psycho_maniaci take it that topic had the word "spaz" in it?
08:06:11 Quit jepler (Remote closed the connection)
08:06:16 Join jepler [0] (n=jepler@emc/developer/jepler)
08:06:16LinusNmy guess too
08:06:26 Join iamben_ [0] (
08:08:53*amiconn wonders why there is an actual op named 'Chanserv' here now
08:10:22scorcheit is a set option for freenode...the upside of which, means that we will not lose modes, banlists, exception lists, etc in event of service failure...this has happened twice before in recent memory in times of freenode chaos
08:12:26 Quit OlivierBorowski (Read error: 110 (Connection timed out))
08:13:40 Join CaptainSquid [0] (
08:18:18 Quit Mouser_X ("Java user signed off")
08:21:03amiconnjhMikeS: The current 1st/2nd Gen build has that crash problem again. First the cop crashes (showing the UIE screen), but rockbox continues too boot. A few seconds later the cpu crashes too...
08:21:33 Quit iamben (Read error: 110 (Connection timed out))
08:25:14 Join Buschel [0] (
08:25:33 Part safetydan
08:28:56Buschelif I play videos via mpegplayer without frame-dropping and frame-limitation (test configuration) I see much higher fps than in real life (frame-dropping and fps-limitation). e.g. my full screen videos play @16fps in test configuration and @12ps with audio/video synced.
08:29:30Buschelis this normal behaviour / technical limitation?
08:30:52 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
08:31:47amiconnI think that even the skipped frames need to be decoded to some degree
08:33:41 Join Mouser_X [0] (n=mouser_x@
08:36:21Buschelyes, but i _all_ frames are decoded i get 20fps, with skipping only 16fps.
08:36:22*LinusN writes another post on the motorola forum
08:37:13 Join Rob222241 [0] (
08:37:29Buschel*ahem all->16fps, skip->12fps
08:37:39Buscheli should drink another coffee...
08:39:10amiconnYes, because the 12 fps with skipping are only the frames shown; the number of source frames is > 16 per second then (probably 24)
08:41:36Buschelhmm, i'm wondering how the sync-mechanism audio<->video is working. maybe i'll take a look at it. but for now i better go working now :)
08:41:40 Quit Buschel ()
08:55:08 Quit Rob2222 (Read error: 110 (Connection timed out))
08:55:27 Quit l7 ("Lost terminal")
08:56:49 Quit lucas42 ("Leaving.")
09:01:35 Quit _pill ("changing servers")
09:02:02 Join reppy [0] (
09:02:35 Join pill [0] (
09:03:35 Join petur [0] (n=petur@rockbox/developer/petur)
09:04:59 Quit hannesd (
09:05:22*amiconn just had an idea
09:05:25amiconnjhMikeS: ping
09:07:12 Join Zagor [0] (
09:07:23 Join lee-qid [0] (
09:08:43*jhMikeS goes and has a smoke
09:09:20amiconnYou said those weird crashes on PP5002 look like a cache issue
09:09:35amiconnDid you test things in the bootloader?
09:10:13amiconnIirc there is a comment in the bootloader saying that we can't enable the cache or else rockbox won't boot - but what if the cache is already enabled by the apple loader on 1st..3rd Gen?
09:12:48 Join spiorf [0] (
09:13:23jhMikeSI'm using a bootloader built from the committed code
09:13:53***Saving seen data "./dancer.seen"
09:14:38amiconnDoes that bootloader handle the cache?
09:14:42jhMikeSIf does the flush/invalidate sequence. Nothing should need saving from the COP cache so it simply invalidates.
09:14:53 Join hannesd_ [0] (
09:14:57amiconnMaybe I should try that too
09:15:18amiconnAn svn build crashes spectacularly on my 1st and 2nd Gen :(
09:15:51jhMikeSI'm not sure it's a cache issue. I notice the dump comes before the rockbox logo and the screen is offset to the left.
09:15:51 Quit scorche (Read error: 104 (Connection reset by peer))
09:16:20 Join ender` [0] (
09:16:45 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
09:17:13jhMikeSWhen it runs, it runs fine. What struck me is that an SVN build was running, then proceeded to stop running at a later time.
09:18:14amiconnYes, when it runs, it runs fine, but now with dualcore, if it crashes, both cores crash...
09:18:22amiconn(cop first)
09:18:59jhMikeSmakes sense since COP will proceed out of kernel init first after waking the CPU.
09:19:33jhMikeSit just stays asleep inside switch_thread
09:21:13 Quit psycho_maniac (" HydraIRC -> <- Would you like to know more?")
09:21:37jhMikeSone test revealed no change if memset was used to zero thread and core structures
09:22:47jhMikeSI think something is just trashing the scheduler data
09:24:11jhMikeSor some object somewhere
09:25:49 Join stripwax [0] (
09:27:33 Quit reppy ()
09:33:51 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
09:41:34 Join hasmind [0] (n=hasmind@
09:41:36 Quit lee-qid (Read error: 110 (Connection timed out))
09:43:32 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
09:45:14 Nick linuxstb__ is now known as linuxstb (
09:50:36 Join davina [0] (
09:52:41 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:59:26 Join _quester [0] (i=c27f0814@gateway/web/cgi-irc/
09:59:52 Quit Ebert ()
10:00:36 Quit TMM ("Ex-Chat")
10:00:46_questerLinusN: hello LinusN. Could you please answer the questions of yesterday 17:35? (Should I repost them now?)
10:02:59linuxstb_quester: Why are you only asking LinusN?
10:06:22 Quit hasmind ("Leaving")
10:06:26_questerlinuxstb: because he made the change and thus must have had a reason to do it that way.
10:07:59 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:08:43_questerBut I'd be of course grateful if someone else would answer the questions :-)
10:09:35petur_quester: it would be easier if you would ask the question again
10:13:12_questerpetur: ok, here we go. I just wasn't sure if a repost is a violation of the irc rules.
10:13:16_questerWhy was it necessary to introduce INCREPEAT? Couldn't the last field just be changed to BUTTON_LEFT (for DECREPEAT for keymap-x5.c)? Since in the end the two actions do exactly the same. So why the difference on the action level? Or do I miss something?
10:15:31linuxstb_quester: I assume you're talking about this commit?
10:17:43_questerlinuxstb: yes
10:19:39safetydanIf I was to unify some of the number formatting functions, where would a could place for it be in the code?
10:20:02safetydanAt the moment, recording.c, settings_list.c and option_select.c all have their own version of the number format code
10:20:12safetydanthere's probably others, but those are the ones I've noticed
10:20:40peturI think there is some code in misc.c already
10:23:00safetydanhrm, not seeing anything exactly like this. There's output_dyn_value but that's not quite right for this.
10:23:10safetydanbut seems like a good place to put this sort of code
10:23:46peturoutput_dyn_value is what I was thinking of
10:37:01 Nick parafin|away is now known as parafin (
10:39:02amiconn_quester: INCREPEAT/DECREPEAT aren't new actions, they already exist in other context
10:39:22amiconnAnd I think that they should be used like that commit does, for consistency
10:43:58LinusN_quester: i can't really say why we have different actions for dec and decrepeat, but since we have them, i decided to make use of them in a consistent manner
10:45:33amiconnThe reason for the separate repeat events is rather obvious
10:45:47amiconnIt's the wrap suppression in lists
10:46:21LinusNah of courese
10:47:02_questeramiconn: ah, then I understand! Thanks! So repeat is essentially the same as many single actions but different with respect to list wrapping.
10:47:43_questeramiconn: hrm... but OTOH isn't wrapping handled at a lower level?
10:48:05_questeramiconn: i.e. sync list level?
10:48:21LinusN_quester: not in the recording screen
10:48:31LinusNwhich was the bug i fixed
10:48:53 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
10:49:08LinusNanmother solution could have been to invent new actions for the gain adjustment in the recording screen
10:49:22LinusNinstead of reusing the setting actions
10:49:31 Nick OlivierBorowski_ is now known as OlivierBorowski (
10:50:15 Quit OlivierBorowski (Remote closed the connection)
10:52:08_questerLinusN: ok. That's another question. My concern was understanding the keymaps code :-)
10:52:20_questerThanks and bye
10:52:22 Part _quester
10:54:15 Quit linuxstb (Nick collision from services.)
10:54:20 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
10:55:52 Join OlivierBorowski [0] (
10:56:32 Quit GodEater_ (Remote closed the connection)
11:01:47 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
11:13:56***Saving seen data "./dancer.seen"
11:14:15 Quit GodEater (Read error: 113 (No route to host))
11:15:06pondlifeHmm, so time for another cup of coffee at AMS, maybe...
11:15:25pondlifeWho visited them last time round?
11:15:32Bagderrp and austriancoder
11:18:38 Quit advcomp2019 (Read error: 110 (Connection timed out))
11:23:40 Quit Soap_ (Read error: 104 (Connection reset by peer))
11:24:15 Join Soap [0] (n=Soap@rockbox/staff/soap)
11:35:06 Quit crashd_ (Connection reset by peer)
11:38:43 Join crashd_ [0] (
11:39:43 Nick Soap is now known as Soap_ (n=Soap@rockbox/staff/soap)
11:50:04amiconnpondlife: That 'access denied' error in voicebox is because the script fails to write its log. The script needs write permissions in its own directory
12:02:01 Join Siku [0] (
12:02:46safetydanbuh? red ipod builds?
12:03:05safetydanoh, disk full on the build server
12:04:22 Join J3TC- [0] (
12:07:46safetydanwoo, 200 bytes saved by moving the units out of the sound drivers
12:08:58 Join Bagder_ [0] (
12:09:58safetydanhave to finish later I guess
12:10:00 Quit safetydan ("Leaving")
12:14:11 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:16:09 Join FOAD_ [0] (
12:19:52 Join f4de [0] (
12:20:00f4dehey i was wondering if i can ask a question?
12:20:32FOAD_Apparently you can.
12:21:30f4deI have a ipod nano i was replacing with rockbox
12:21:33f4deand i was just wondering
12:21:41f4deit comes up with found new hardware and says rockbox player
12:21:46f4deand i havent got any drivers for it
12:22:21 Join Thundercloud [0] (
12:22:22 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
12:22:54FOAD_What generation?
12:23:21f4dei used the QT program
12:23:24f4dewas working all fine
12:23:34f4detill i took it out to reboot got rockbox on it
12:23:45f4debut now comes up with add new hardware instead now
12:24:22rasherf4de: rockbox' USB support is in very early stages. You should enable diskmode or reboot to the Apple firmware to connect
12:25:29 Quit Bagder (Read error: 110 (Connection timed out))
12:25:31f4debut if i cant connect to it
12:25:38f4dehow can i reboot to the apple firmware?
12:26:12rasherYou hold some button-combination.. Someone with an ipod should tell you - I don't know the details.
12:26:23f4deyeah im just doing it for my mum
12:27:15Zagorf4de: ignore that driver popup.
12:27:20f4dei just cancelled and it let me in
12:27:22f4dethanks :)
12:27:47Zagorit's a temporary kludge to enable full-power charging in rockbox
12:30:28f4deoh thanks alot zagor :)
12:31:48 Quit J3TC- (Read error: 113 (No route to host))
12:33:15 Quit FOAD (Read error: 110 (Connection timed out))
12:33:15 Nick FOAD_ is now known as FOAD (
12:33:39jhMikeSamiconn: Cache problem it seems after all. Now that I can run it with the cache off correctly, even the battery level is good from startup. Next thing is a proper fix I suppose whatever that end up as.
12:33:59 Quit linuxstb (Read error: 110 (Connection timed out))
12:38:06 Quit f4de (" HydraIRC -> <- Chicks dig it")
12:38:39 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
12:42:37 Nick billenium is now known as billeniumZzZ (
12:46:17 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
12:48:37 Quit linuxstb (Nick collision from services.)
12:48:38 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
12:51:05 Join homielowe_ [0] (
12:51:44pixelmaZagor, Bagder_: one of you already looked into the issue? Thought I'd ask in case you weren't aware...
12:52:20Zagorwhat issue is that?
12:53:27pixelmathe build table shows failed builds on that one, seems to be no space left on device (see failed M5 sim build a bit down in the table)
12:54:54pixelmathe red in the last build round is also on that server but the error message is a bit confusing
12:56:02 Join rp- [0] (n=rp@
12:57:52preglowlooked like /var has a tad too little space
12:58:54Zagorit's only 19G, which is obviously too small :)
13:00:09preglowi wonder why sdoyon modified rbspeexenc to have a volume flag when espeak itself has that...
13:01:13amiconnFor symmetry, I think.
13:01:36amiconnWith mp3, some people used the lame −−scale parameter
13:01:48amiconnNot all tts systems have a volume setting
13:02:47 Join barrywardell [0] (
13:03:50 Join iamben [0] (
13:05:03 Join stewball`ghost [0] (n=WTFOMGBB@
13:14:00***Saving seen data "./dancer.seen"
13:14:48 Quit iamben_ (Read error: 110 (Connection timed out))
13:15:33preglowwhen you set a wake-up alarm, you should be able to switch the unit on and off as many times as you want, and still have it wake up after you finally turn it off, yes?
13:16:05linuxstbSounds reasonable...
13:16:45preglowwell, ipods don't do that now
13:17:07preglowand i don't really know how to solve it, i had to nerf that when i fixed the "mysteriously waking ipods" problem
13:17:24amiconnThat would need additional handling, making sure the alarm doesn't fire while rockbox is running
13:18:25preglowwhat's the harm in that?
13:18:48markunZagor: do you think firmware/usb_serial.c can easily be converted to work with the new USB stack?
13:20:37ZagorI think it would probably be both easier and better to write a new one using that as documentation/inspiration. first it's got all the driver details mixed into the protocol handling, and second it is for a full-speed driver which work very different from high-speed ones
13:22:22Zagorhowever as serial typically uses only very small packages, it is not unlikely that we could get serial to work even without solving the big block bug
13:22:37markunthe Gigabeat F is only full-speed. Does that mean it will be difficult to port your USB stack?
13:23:49Zagorno I don't think so
13:24:25preglowfull-speed is what?
13:24:30preglowmore than 1.1, i hope? :>
13:24:49Zagorthe main difference is that on high-speed devices you prepare the controller to respond to requests before they are sent by the host. full speed is much more linear request/response.
13:25:04Zagorfull speed is 12 Mbit/s, high speed is 480 Mbit/s
13:25:08amiconnZagor: Btw, does that special array in the usb driver really need 2048-byte alignment
13:25:24 Quit homielowe_ (Remote closed the connection)
13:25:36 Quit Zom (Read error: 110 (Connection timed out))
13:26:05Zagoramiconn: yes it does. or at least the manual says so. I don't remember if I've tried it unaligned.
13:26:29 Join J3TC- [0] (
13:30:32Zagoralso if you want to use the full 20K buffer capability, that too must be aligned to 2K... it's something to do with using very few bits for index counters, I think.
13:33:14 Join Zom [0] (
13:34:17Nico_Phaha unzooming plasam a lot can produce interesting effects
13:36:15 Nick fxb__ is now known as fxb (
13:38:35 Join Llorean [0] (
13:39:48 Join aliask [0] (
13:42:22Nico_Paliask: hi
13:42:42aliaskHowdy Nico_P
13:42:46 Quit atsea-104 (Remote closed the connection)
13:42:56Nico_Pthanks for editing the wiki page :)
13:43:31aliaskI hope I didn't miss anything. Would a cross compiled sendfirm file work on windows?
13:43:55Nico_PI don't think so, as libmtp is only for POSIX systems AFAIK
13:44:24aliaskA shame.
13:45:00Nico_Phave you tried anything new?
13:45:28aliaskI began looking at disassembling the OF bootloader in IDA
13:45:38aliaskI didn't get very far though
13:46:39 Join atsea-104 [0] (i=atsea-@gateway/tor/x-92b3ec3fde81b319)
13:47:33Nico_Pdo you thinkw e could attempt to load the firmware from a file?
13:51:23linuxstbYou need to change the DRAM start address in firmware/ to 0x88000000 and then load the firmware binary to that address, and branch to it.
13:51:24Nico_POK, I'll start by getting the main binary to compile
13:51:37aliaskI had some thoughts on how to ID the file, it'd be slow, but work. The rockbox.gigabeat file has an ID at the top, the first four bytes would be "gigs", so we could scan files larger than x kb for that string
13:51:50Nico_Plinuxstb: I guess I can look at the dual boot code for inspiration?
13:52:07 Quit atsea-104 (Remote closed the connection)
13:52:08preglowdo any of you people know if fopen() and co expects a filename with the system locale encoding, or something else?
13:52:11linuxstbNico_P: Probably the other ARM bootloaders would be better.
13:52:22Nico_Plinuxstb: ok
13:53:12Nico_Paliask: yes, that and it also appears that the firmware file doesn't have an extension
13:53:32Nico_Pand maybe for now we can assume it'll be 02 too
13:53:49aliaskNico_P: We can't assume it'd be 02 for a player which already has media on it
13:54:10Nico_Pyes but I think for testing I'll start by making that assumption
13:54:19aliaskThat sounds reasonably
13:54:37LloreanWill we be able to have our own non-MTP USB mode eventually?
13:54:48linuxstbI certainly hope so...
13:54:58Nico_Pprobably with Zagor's USB stack
13:55:01LloreanIf so, you should never have to make an assumption. Just have them install a bootloader with USB, then do a normal USB connection and copy the build over
13:55:13LloreanWell "never after things work"
13:55:24LloreanSo it's a purely software USB device?
13:55:29Nico_PLlorean: yes, that's what I was thinking too
13:56:33LloreanBesides, a build can't really work if the rest of it doesn't have the filenames Rockbox wants...
13:57:52aliaskNico_P: Could I have a binary of the libmtp sendfile, I can't manage to compile mine - complains about a missing config.h
13:57:57ZagorI could really use some help from some of the sharp minds (and eyes) here.
13:58:17Nico_Paliask: you need the whole source I think, but yeah I can send you the bin
13:58:39aliaskI do have the whole source, there's no config.h file in there.
13:58:41linuxstbaliask: You're trying to compile the version in SVN?
13:59:11aliask0.2.2 tarball
13:59:24Zagorbasically my driver code needs to be closely scrutinized, side by side with the usb controller reference manual. there is a bug in there that's preventing big blocks from working. and I'm not finding it.
13:59:34linuxstbaliask: Sorry, I thought you meant sendfirm...
13:59:52Nico_Paliask: have you run make in the main src dir?
14:00:14aliaskYep. I have libmtp installed from there
14:00:49linuxstbaliask: Did you run ./configure ?
14:02:19aliaskHrm, I don't remember. I'll try rebuilding again
14:02:42pixelmasomeone with Archos knowledge read today's log? There was someone describing weird behaviour on an FM Recorder (starting at around 5 o'clock) with the latest builds, the details look "interesting"...
14:04:08linuxstbNico_P: Do you get a "sendfile" executable, or is it part of "connect" ?
14:04:29Nico_Plinuxstb: it's part of connect, with a symlink pointing to it
14:04:52Nico_Pmost of the example programs actually use connect
14:05:02linuxstbHave you tried creating new folders?
14:05:05barrywardellZagor: where do you think the bug is likely to be?
14:05:05LloreanZagor: Current code is in SVN, and the manual is in the wiki somewhere? (More for the logs, I find it somewhat unlikely I'll be able to spot something)
14:05:16Nico_Plinuxstb: on the device?
14:05:19 Join MethoS- [0] (
14:05:31linuxstbNico_P: Using "connect −−newfolder [foldername]"
14:05:43linuxstbNico_P: And then putting files in there?
14:05:47Nico_Pno, I'll try
14:06:15aliaskAh, connect compiled for me. I'll just use that
14:08:15Nico_Plinuxstb: mtp-newfolder reported success, but mtp-folders doesn't show anything
14:08:29 Quit MethoS- (Remote closed the connection)
14:09:30Nico_PI sent a file to the device, asking to put it in the folder I tried to create... when loading the rockbox bootloader the file appeared in the same place as the other files
14:10:07Nico_PI got the main binary to compile btw
14:12:05 Join MethoS- [0] (
14:12:12Zagorbarrywardell: I think it's most likely in the usb-drv-pp502x.c driver. init_queue_heads(), init_endpoints() and prime_transfer() are prime candidates
14:14:01 Join PaulJam [0] (
14:14:10aliaskNico_P: When connect wants a destination, what do you give it?
14:14:21Nico_Paliask: just /
14:14:31Nico_Pit's the destination dir
14:16:17ZagorLlorean: yes, the source is in svn: and the controller reference manual is here:
14:17:13linuxstbNico_P: I think adding something like asm volatile("mov pc, #0x88000000"); to the end of bootloader/gigabeat-s.c will do what we want - after you load the Rockbox binary to that address.
14:18:20linuxstbNico_P: And don't forget to change 0x80000000 to 0x88000000 in firmware/ (DRAMORIG for IMX31L)
14:18:41Nico_Plinuxstb: ok. for now I'm checking that I'm able to read the file correctly
14:19:46 Join GodEater [0] (i=c2cbc95c@rockbox/staff/GodEater)
14:22:46Nico_PI'm having doubts
14:23:16 Quit Febs (Read error: 110 (Connection timed out))
14:25:43barrywardellZagor: remind me...RX and TX both don't work with big blocks?
14:26:36 Join hannesd__ [0] (
14:28:00Nico_Plinuxstb, aliask: unless I'm doing something stupid, the first bytes of what I'm reading don't match with the first bytes of the file I sent
14:28:18aliaskFilesize is correct?
14:28:42 Quit JdGordon ("Konversation terminated!")
14:28:55aliaskIt could be that ATA isn't really working properly
14:29:20 Quit homielowe (Read error: 110 (Connection timed out))
14:31:29 Quit spiorf (Remote closed the connection)
14:31:56linuxstbMaybe the OF is encrypting the files...
14:32:50Nico_Phmm actually read is returning -1, I had forgotten to check the return value
14:33:32 Join OlivierBorowski_ [0] (
14:38:03*Nico_P slaps forehead
14:38:53Nico_PI should've considered the "I'm doing something stupid" option more seriously... I was closing the file before reading it
14:39:08Nico_Pnow it's fine
14:39:28barrywardellZagor: what do I need to do to make try MSC mode in Rockbox?
14:40:32Zagorbarrywardell: change at the top of usbstack/usb_core.c to define USB_STORAGE instead of USB_CHARGING_ONLY
14:40:49barrywardellok, thanks
14:41:30barrywardellbtw. rebooting on usb connect isn't currently working
14:41:38Zagoron which player?
14:41:58 Quit hannesd_ (Read error: 110 (Connection timed out))
14:42:04Zagordo you have bagders commit from late yesterday?
14:42:49 Quit GodEater ("CGI:IRC (Ping timeout)")
14:43:05barrywardellI just did svn up, r15833
14:43:21 Join atsea- [0] (i=atsea-@gateway/tor/x-c3e34d7edfbf7117)
14:43:28Zagorok. does it crash or does it just not respond?
14:43:34Nico_Plinuxstb: can't I use kernel_entry to load the firmware?
14:43:37Zagorjust not reboot
14:44:11barrywardellsometimes crashes, sometimes the screen goes black and the wheel light stays on and nothing happens
14:44:32Nico_Plinuxstb: basically, just reenable the code that is in #if 0
14:45:12Zagorbarrywardell: ok. bagder said his commit fixed that issue on his e200.
14:45:15linuxstbNico_P: You would need to look in crt0.S to see what it does...
14:45:36Zagorbarrywardell: apparently more investigation is needed to get a safe system_reboot() for e200.
14:46:03 Quit J3TC- (Read error: 110 (Connection timed out))
14:46:29 Join J3TC- [0] (
14:47:26linuxstbNico_P: No, I don't think you can - the crt0.S for the S doesn't support that.
14:48:07Nico_Plinuxstb: where is it?
14:48:11aliasklinuxstb: Out of interest, why not?
14:48:15Nico_Pand more importantly, what is it?
14:48:46Nico_Pah, target/arm/crt0.S ?
14:49:08aliaskNico_P: That's the one
14:49:27aliaskIt's the assembly code that gets run at the very start of the bootloader
14:49:42preglowamiconn: if the english.voice, english-1.voice, english-2.voice thing was very intentional, sdoyon just broke it
14:49:42aliask(and also the main build, in the appropriate sections)
14:50:14 Join GodEater [0] (i=c2cbc95c@rockbox/staff/GodEater)
14:51:10linuxstbaliask, Nico_P: Forget what I said, I was confused. That kernel_entry() thing has nothing to do with crt0.S, so should work fine.
14:51:20 Quit OlivierBorowski (Read error: 110 (Connection timed out))
14:52:01 Quit axionix (Read error: 104 (Connection reset by peer))
14:52:22linuxstbNico_P: You'll need to add "int(*kernel_entry)(void);" to the start of main() in bootloader/gigabeat-s.c (see bootloader/gigabeat.c)
14:53:04 Join axionix [0] (
14:53:59Nico_Pwow it kinda worked
14:54:36Nico_PI got the rockbox boot screen (logo and version), but with a weird yellow line and it's frozen
14:54:45aliaskI wonder why it never worked for me...
14:55:21linuxstbWhere were you loading it to?
14:56:08aliaskProbably some really wrong address, and I probably left the pointing to 0x80000000
14:56:17aliaskAnyway, that's great news Nico_P
14:57:12 Join japc [0] (n=japc@
14:57:19Nico_PI guess a patch is in order
14:57:42 Join spiorf [0] (
15:03:40san|chezanyword on the 3G ipod firmware being fixed at all?
15:04:04 Quit Siku (Nick collision from services.)
15:04:10 Join Siku [0] (
15:05:05 Join Febs [0] (n=chatzill@
15:06:14Nico_Paliask, linuxstb: the checksum seems to fail... things were loading because I had removed the check
15:07:37linuxstbsan|chez: There was a commit last night - see (second to last entry)
15:07:49Nico_Phmm yeah the logo being displayed seems to have been pure luuck
15:08:20san|chezthanks linuxstb
15:09:14 Join MethoS-- [0] (
15:09:14san|chezi was really upset when i have to uninstall rockbox 10 minutes after flashing :)
15:09:20*san|chez cant stand iTunes
15:14:01***Saving seen data "./dancer.seen"
15:14:44 Quit Siku (
15:14:44 Quit rasher (
15:14:44 Quit Slasheri (
15:14:44 Quit Jon-Kha (
15:14:44 Quit Hadaka (
15:14:56amiconnpreglow: rasher wrote that, and it was intentional as far as it was meant to allow for building several voice files (with different voices, tts parameters etc) in one go
15:15:02amiconnBut that never worked afaik
15:17:11aliaskNico_P: Sounds like ATA issues to me...
15:17:29 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
15:17:31Nico_Paliask: I'm not sure... the checksum is really close to what it should be
15:17:55aliaskWhere is it missing I wonder?
15:18:02 Join rasher [0] (
15:18:13Nico_Pand it appears that the file is read correctly, but the bootloader is checking against the wrong checksum
15:18:36aliaskWhat do you mean?
15:19:28Nico_Phmm no I think I hadn't understood how the check is done
15:20:33preglowamiconn: well, i've been confused several times, using wrong voice files, so i don't mourn seeing that feature go, heh
15:20:55 Nick fxb is now known as fxb__ (
15:21:09 Join Naked [0] (
15:21:09NJoinSiku [0] (
15:21:09NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
15:21:10 Join Jon-Kha [0] (
15:21:10Nico_Paliask: where does gigabeat_s_code come from?
15:21:12***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
15:21:13 Nick Naked is now known as Hadaka (
15:21:51aliaskIn scramble.c?
15:22:25Nico_Paliask: in tools/gigabeats.c... did you write it?
15:22:38linuxstbIt looks like MODEL_NUM is wrong in firmware/export/config-gigabeat-s.h - it should be 21 (to match tools/scramble.c)
15:22:56linuxstbNico_P: Is the checksum wrong by 2?
15:23:04Nico_Plinuxstb: yes
15:23:06aliaskI did. I'm actually working on it right now. Most of it's copied from the OF nk.bin in an attempt to get the gigabeat V updater to accept it
15:23:15linuxstbNico_P: That's the problem then.
15:23:21Nico_Plinuxstb: thanks :)
15:23:51aliaskBut I'm simplifying it in the hope that sendfile will get it there and it'll work anyway
15:23:57 Quit Siku (Read error: 145 (Connection timed out))
15:24:05linuxstbNico_P: "tool" is set to be "scramble -add=gigs" in tools/configure - this is the standard 8-byte header consisting of a 32-bit checksum and four character model name.
15:24:54linuxstbNico_P: The gigabeat_s_code() function is used when calling "scramble -gigabeats"
15:25:23Nico_Pyes, I was looking at it and didn't understand what it does
15:25:27Nico_Pnow I get a data abort
15:26:45Nico_Pat 80040DC0
15:27:03linuxstbSomething is still wrong then - there shouldn't be any code there...
15:27:11 Quit MethoS- (Read error: 110 (Connection timed out))
15:27:16linuxstbYou changed firmware/ ?
15:27:54Nico_Plinuxstb: yes, I changed 0x80000000 to 0x88000000 for DRAMORIG
15:28:26linuxstbDo you get it immediately Rockbox loads?
15:30:06Nico_Plinuxstb: I randomly set buffer_size to 1024*1024 for load_firmware
15:30:40Nico_Plinuxstb: does an change warrant a reconfigure?
15:30:54linuxstbNo, it shouldn't do, it's just used at the final linking stage.
15:31:02aliaskI just ran a generated nk.bin file
15:31:25Nico_Paliask: no OF in it?
15:31:38linuxstbSo just two chunks - the eboot disable and the bootloader?
15:31:44Nico_Pice :)
15:31:57aliaskThree. One for 0x88000000
15:32:12aliaskldr pc, =0x8a000000
15:32:40aliaskIt supprised me, I thought something had gone wrong because it loaded so fast
15:32:46linuxstbDid you try with just two, and setting the entry point to 0x8a000000 ?
15:32:58linuxstbI can imagine - there's 13MB of OF to load normally...
15:32:59aliaskNope. I can do that now though.
15:33:24linuxstbBut that's good - it means we can generate a clean nk.bin we can distribute.
15:33:34aliaskBut windows users are left in the dark
15:33:41aliaskThe gigabeat V updater won't send it
15:33:47linuxstbLet's not worry about them yet...
15:34:18 Quit jhMikeS (Nick collision from services.)
15:34:24 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
15:34:36aliaskWithout the OF we must use the recovery mode to update the bootloader unfortunately.
15:34:37linuxstbI wonder if there is an official win32 API for MTP that acts in a similar way to libmtp...
15:35:04Lloreanlinuxstb: Can windows users use vmware?
15:35:13Nico_PI guess we could ask the libmtp devs about that... they ought to know
15:35:13linuxstbLlorean: You tell us... ;)
15:35:38LloreanAt the moment I'd have to load a vmware image of debian within a vmware image of windows, not a fair test. ;)
15:35:51markunlinuxstb: maybe this can help
15:35:57Nico_Plinuxstb: about the change, I'm not sure I recompiled my binary after making the change, I'll do it again
15:36:15aliasklinuxstb: Two records works fine
15:36:32linuxstbaliask: Nice...
15:36:43linuxstbmarkun: Looks useful...
15:37:05linuxstbmarkun: Nice indenting here ;)
15:38:37*markun wonder what he's missing..
15:38:49Zagormarkun: it's rather ... wide
15:38:58aliaskNo 80 column guideline there...
15:39:03Nico_Phow do you fit that in 80 columns?
15:39:23linuxstbYou don't have 40 nested if() statements...
15:39:44markunoops, didn't see that file :)
15:39:53 Quit spiorf (Remote closed the connection)
15:40:12ZagorI love the scattered little "else" lines at the bottom.
15:40:53markunDid anyone listen to the test file?
15:41:18markunprobably not terribly interesting
15:41:29markuncompared to "catch the digital flow" for example ;)
15:42:05Lloreanpreglow: rbxspeexenc just creates speex data without a container? Am I remembering correctly?
15:43:18 Join n1s [0] (
15:43:40Zagorhmm, usb works slightly better in full speed mode
15:43:59LloreanSlightly better?
15:44:16Zagoryeah I can send one full 16KB buffer, and then it fails at the second
15:44:35Zagoror I can send a dozen 512 KB buffers
15:44:47Zagorin high speed, it always failed on the first buffer I try to send
15:44:54 Quit rp- ("leaving")
15:45:38Nico_Plinuxstb: now I'm back to getting a slightly corrupted splash screen
15:46:18preglowLlorean: yes
15:46:24preglowLlorean: i'm going to elaborate on the ml
15:46:41Nico_Plinuxstb: what is the entry point?
15:47:04Zagoralso in full speed I don't have any problems with receiving. 102400 bytes in 103 ms = 994 kB/s
15:47:11linuxstbNico_P: i.e. the first instruction in the binary.
15:47:21preglowLlorean: for voice files, the voice file itself is a container. using proper .ogg files for talk clips would be nice, but would bloat the core
15:47:26Nico_Plinuxstb: yes, but what's there once I've loaded the binary there? ie where in the source is the first intruction? main()?
15:47:35aliasklinuxstb: Shouldn't it be 0x88000008
15:47:47linuxstbNico_P: The first instruction is in firmware/target/arm/crt0.S
15:47:50linuxstbaliask: Why?
15:47:50 Join nanok [0] (n=nanok@
15:48:04aliask4 bytes for the "gigs", and 4 bytes for the csum
15:48:09linuxstbNico_P: Which then calls main() in apps/main.c
15:48:13linuxstbaliask: You don't load those 8 bytes.
15:48:19 Nick Bagder_ is now known as Bagder (
15:48:23Nico_Plinuxstb: ok, thanks
15:48:31aliaskOk. I misunderstood
15:48:33linuxstbaliask: Or rather, you load them separately to check, and load the rest of the file to 0x88000000
15:49:30linuxstbaliask: Do you know what is happening with interrupts and memory remapping in both the bootloader and main firmware?
15:49:43linuxstbi.e. are interrupts enabled, and is any memory being remapped?
15:50:23aliasklinuxstb: Memory is certianly being remapped. DRAM is supposed to be 0x80000000 according to the datasheet.
15:50:43linuxstbWho wrote that code in crt0.S?
15:50:44aliaskBut no idea about anything else.
15:51:00 Quit animeloe (No route to host)
15:51:04aliaskptw419 did. But I think it was taken from the linux bsp
15:52:35aliasklinuxstb: Should I commit my work on generating the nk.bin file?
15:54:32 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
15:54:57 Quit CaptainSquid ("Miranda IM!")
15:55:34Nico_Paliask, linuxstb:
15:56:50linuxstbaliask: Sure. A couple of things I noticed from the code in SVN - it's not 64-bit safe (use of longs) and it's not endian-safe.
15:57:04markunlinuxstb: I see that the Cowon iaudio A3 plays APE, do you know which compression level?
15:57:18linuxstbmarkun: No idea...
15:57:24aliasklinuxstb: I'll borrow some code from mknkboot to fix those up before I commit.
15:58:00markunIt's powered by the TI Davinci
15:58:29Bagdermarkun: how powerful arm does that have?
15:58:44linuxstbmarkun: There was a post recently to the flac-dev mailing list advertising the flac capabilities of the A3 - it supports FLAC in .flac, .ogg and .mkv and also supports realtime FLAC encoding from line-in.
15:59:14GodEatermkv? Wow - you don't see that too often!
15:59:21linuxstbThat's what I thought...
15:59:45GodEaterthat's matroska I assume ?
15:59:56Bagderok, ARM9 up to some 270MHz
15:59:56GodEaterwonder what the "v" is for
16:00:02LloreanI thought "mkv" was specifically Matroska Video...
16:00:23linuxstbThe email said "mkv/mka" but I was lazy...
16:00:31LloreanAh, okay.
16:00:47linuxstbI guess there is no difference - just a convention for the filenames depending on how many streams there are in the file, and what types they are.
16:01:26linuxstbSimilar to the recent change to use "oga" for audio-only Ogg streams, and ogv for audio/video.
16:01:45Bagderand the tms320dm355 is... *publicly documented*
16:01:47preglowyeah, the finally realized how it should have been in the first place...
16:02:20ZagorBagder: yay
16:02:22Bagderman, I've started to forget how the times were when our targets had docs
16:06:54markunI wonder what made them change their mind
16:07:23Bagderyes, and why they then didn't release the docs for the other models too...
16:08:00markunYes, why don't the same reasons they used to have apply to this processor..
16:08:35markunmayse someone in the company got replaced
16:09:02markuncould be the same reason they now let us use their compiler
16:09:52 Quit Zagor ("Client exiting")
16:13:02linuxstbNico_P: I've just looked at your patch - how did you choose the filename (/Content/0B/0000/05) ?
16:13:42Nico_Plinuxstb: from the recursive dir listing I have... I'm chaging things so that the first firmware file is loaded
16:15:15linuxstbSo it searches for a file with "gigs" in the header?
16:15:27Nico_Pthat's what I'm writing now
16:15:35linuxstbOK, I was about to suggest you did that ;)
16:17:02GodEaterhehe - reminds me of that Penny Arcade cartoon with "Gigs" in it
16:18:15linuxstbGodEater: ?
16:18:33GodEateryep - that one ;)
16:30:53Nico_Plinuxstb: I'd like to commit but I need to load the second partition... I hack disk.c to do that but I can't commit in that state
16:31:23aliask#ifdef GIGABEAT_S?
16:31:54GodEateris the plan to support the two parition layout forever ?
16:32:05Nico_PGodEater: I doubt it
16:32:21GodEaterthen yeah, #ifdef it temporarily
16:32:29GodEaterwith a view to removing it further down the line
16:33:44*n1s is happy to see nice gigabeat s progress :)
16:34:20 Quit Thundercloud (Read error: 104 (Connection reset by peer))
16:35:28*linuxstb tries to tempt GodEater and n1s to join the S-fun...
16:36:07n1slinuxstb: my lowlevel skills are not quite there yet...
16:36:22GodEaterlinuxstb: my home PC died properly today - I have no development platform :(
16:36:43linuxstbGodEater: You have a ps3 ;)
16:36:54GodEaterhowever, I have new bits in the post - so will have a whole new PC by the weekend hopefully
16:36:55Nico_Pn1s: mine are very flaky too but I find things to do :)
16:37:03GodEaterlinuxstb: :P
16:37:37GodEaterI can't see Mrs. GodEater being very keen on the TV being covered in source code rather than "I'm a celebrity" going down very well!
16:37:45pondlifeDouble ouch
16:37:54pondlifeYou have that problem too :/
16:38:38pondlife"I'm An Open Source Developer- Let Me Deal With The Bugs!"
16:38:42GodEaterpondlife: I try to ignore it =/
16:39:46GodEaternice play on words
16:43:45 Nick OlivierBorowski_ is now known as OlivierBorowski (
16:44:49 Join Arathis [0] (
16:45:04aliaskWhat's the deal with gcc choking on uint32_t in function arguments and function types?
16:45:26 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
16:46:23 Quit GodEater ("CGI:IRC (Ping timeout)")
16:46:42aliaskWait... gcc isn't letting me declare ANY uint32_t's...
16:47:01n1sinclude <inttypes.h>
16:47:56 Quit atsea- (Remote closed the connection)
16:47:56aliaskBingo. Thanks n1s.
16:49:00 Quit mokkurkalve ("leaving")
16:50:18Nico_Pwhoops... nice yellow line :/
16:55:16amiconnYeah, the stray 'i'
16:55:39Nico_Pamiconn: yes, I just committed the fix
17:01:00 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
17:02:52*pondlife wonders if we'll ever have a fully green build table.
17:03:10pondlifeBlimmin' argument 5...
17:03:16LloreanIt happens more often than eclipses.
17:03:52pondlifeThere's no sensible cast for cfft.c?
17:06:22Nico_Plinuxstb: are you able to do more than one mtp operation on your S without having to disconnect and reconnect?
17:07:20 Join Thundercloud [0] (
17:08:20linuxstbNico_P: I haven't tried. But I don't think MTP is working very well - the device is constantly connecting and disconnecting when I plug it in, and I need to run sendfirm repeatedly until it catches the S connected....
17:09:26 Join webguest10 [0] (i=415fa2c6@gateway/web/cgi-irc/
17:09:34 Quit webguest10 (Client Quit)
17:10:57Nico_Plinuxstb: ah, I used to have that too before, but not anymore... now I need to disconnect between each operation
17:11:52 Join Arathis2 [0] (
17:12:32linuxstbNico_P: How did you stop that?
17:12:39Nico_Pno idea
17:12:59linuxstbpondlife: That seems to be a gcc bug (IIRC)
17:13:36pondlifeIndeed, but I was wondering if some ridiculous casting might get a green build :/
17:13:40 Part LinusN
17:13:49 Join jgarvey [0] (
17:14:02***Saving seen data "./dancer.seen"
17:14:14pondlifeOr some very local way of disabling that warning?
17:22:25*aliask gives up
17:22:29Nico_Plinuxstb, aliask: have you tested the firmware loading?
17:22:37Nico_Paliask: what are you giving up?
17:22:47aliaskThe md5 hashes of two files are identical. One file loads. The other does not.
17:23:26nanokaliask: permissions something?
17:23:27lostlogicaliask: you got linux? try the cmp command
17:23:54aliasklostlogic: Didn't return anything. I'm guessing that means they ARE the same
17:23:58Nico_Plostlogic: that probably won't return anything if the files are identical
17:24:09nanokif md5 are identical and the files not, i think you are the luckyest man ever to have walked this planet
17:24:12lostlogicjust checking, always the 1 in 120398120398120398123098123098123098120398 chance of an md5 collision
17:24:16aliaskNot some freak of nature having different md5s
17:24:21Nico_Plostlogic: btw, do you have any comments on my fix for FS #8092?
17:24:33lostlogicNico_P: that's what I said "Yay" about yesterday ;)
17:24:46Nico_Phaha... apart from that?
17:25:05 Nick hannesd__ is now known as hannesd (
17:25:08lostlogicnothing comes to mind −− it looks good to me.
17:25:09Nico_PI fixed a problem Lear made me notice
17:25:14lostlogicyeah, I saw that
17:27:08lostlogicNico_P: I'll review it carefully later, but conceptually it is exactly what it should be.
17:27:56pondlifeSo, is MoB now complete?
17:27:57Nico_Pcool. I decided to go with a flush instead of making the buffering thread close handles... seemed simpler and it gives the user a chance to decide which handles are most important
17:28:16Nico_Ppondlife: there is still the ID3v1 issue left
17:28:22pondlifeAh, yes
17:28:22lostlogicand cue sheets possibly?
17:28:36Nico_Pyes, that too... I have the start of it
17:28:52lostlogicand I have a plan for id3v1 but haven't started coding yet
17:29:06Nico_PI also want to add the buffering funcs to the plugin API and play with them a bit
17:29:10 Quit Arathis (Read error: 110 (Connection timed out))
17:29:14 Quit axionix ("leaving")
17:29:24 Nick Arathis2 is now known as Arathis (
17:29:26Nico_Paliask: do you want me to try your files?
17:29:27 Join axionix [0] (
17:29:40 Join Frazz [0] (
17:30:03aliaskNico_P: It just started working again
17:30:16aliaskI think I'm just making dumb mistakes because it's 3:30am
17:31:50 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
17:32:03aliaskI'm going to commit this and go to sleep
17:32:09*BigBambi does the S progress dance
17:32:11Nico_Paliask: cool
17:32:19BigBambiCongrats all so far
17:32:42aliaskStill a long way to go yet
17:32:55Nico_Pwe need to figure out how to init everything
17:33:09Nico_PI guess it's time to look at the linux BSP
17:33:15BigBambiYeah, but progress is progress
17:33:18pondlifeNico_P: You brave man, taking on #2687 :)
17:33:26pondlifeThat's two devs it's eaten now
17:33:41Nico_Ppondlife: I think FS #8206 is pretty much the same
17:34:09Nico_Pnote the task names end the same too :)
17:34:43pondlifeYes. Need to make sure all the relevant variables/interfaces are clearly labeled as codec_ or wps_ related transitions..
17:35:03Nico_Ppondlife: the problem is that the WPS always assumes a seek is possible
17:35:09pondlifeWell, it should be
17:35:22pondlifeEven if it needs a rebuffer
17:35:43pondlifeAnd there's a bit of a delay while the codec gets back to work
17:36:01 Join mokkurkalve [0] (
17:36:14pondlifeInternal weaknesses of playback.c are no concern to the UI
17:36:22Nico_Pgood point
17:37:08pondlifeThat's a tricky bug though; I had a hack at it once.
17:37:24pondlifeMight result in further code cleaning ;)
17:37:28Nico_PI guess with low latency it would go away on its own
17:37:39pondlifeNot completely
17:37:44pondlifeBut the window would be very small
17:38:00pondlifeI think we should fix it anyway, it's still a bug
17:40:16aliaskOk, I'm going to sleep. If the build goes all red, blame solar flares.
17:40:20mokkurkalveWhoo-hooo! Got my e280 today. Luckily we're a few weeks behind here in Norway. It's a v1 (and not "Rhapsody" either). Now my rockboxed H340 can get a rockboxed little brother...
17:40:25Nico_Paliask: nice :) good night
17:40:33 Quit aliask ("ChatZilla 0.9.79 [Firefox]")
17:41:29 Quit petur ("*plop*")
17:43:25Nico_Plostlogic: there is one we haven't dealt with... double skipping in cases of codec error
17:45:03 Quit jhulst ("Konversation terminated!")
17:48:14 Quit hcs ("Leaving.")
17:52:22 Quit J3TC- (Read error: 110 (Connection timed out))
17:52:43 Join J3TC- [0] (
17:55:49 Join Domonoky [0] (
17:57:25Nico_Plinuxstb: aliask's commit works nicely and blazingly fast here... not very useful for now but very nice for when we'll have made more pregress
17:58:59lostlogicNico_P: oh, we haven't dealt with that? blah.
18:07:23 Quit pill (Nick collision from services.)
18:07:35 Join pill [0] (
18:08:22 Join spiorf [0] (
18:09:55markunmokkurkalve: congrats!
18:18:45 Part pondlife ("Gone")
18:20:49 Quit axionix ("Lost terminal")
18:21:45 Quit pill (Nick collision from services.)
18:21:57 Join _pill [0] (
18:29:01 Join bertrik [0] (
18:30:03preglowrockbox just played the last track in a dir, then spat a couple of codec failures, then plays the last track again :D
18:30:10*preglow feels there are still bugs
18:30:39Nico_Ppreglow: is auto dir change enabled? what's the repeat mode?
18:30:45preglowauto dir change is indeed enabled
18:31:01preglowrepeat is off
18:31:17preglowauto dir change feels a bit glitchy after mob
18:31:28LloreanWasn't auto dir change always kinda hacky?
18:31:40preglowhacky, but it worked
18:31:43preglownow it's glitchy
18:31:57preglowit failed again when playing the last track, and skipped an entire directory
18:32:07Nico_Pvery weird
18:32:13preglownow it disabled my dircache
18:32:27 Join Ebert [0] (
18:32:52preglowafter the last codec failure, everything spins the disk up
18:33:00Nico_Pit worked nicely here
18:33:17Nico_Ppreglow: does it consistently fail?
18:33:27preglowNico_P: not consistently, no, but often
18:33:42 Quit Ebert (Client Quit)
18:33:51preglowthis isn't the latest build, btw, but it's recent
18:34:07Nico_Phow recent?
18:34:28preglowit always never does what it should, though
18:34:36preglowlikes to skip files and dirs
18:35:01preglowit's from five days ago
18:35:26 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
18:35:54Nico_Pshould be fine then
18:37:39preglowi'd love it if auto dir change was less hacky :/
18:42:06*amiconn would rather remove it :>
18:42:34 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:42:48 Join Fraser [0] (
18:42:52*preglow can't stand using playlists
18:42:54 Part keanu ("Leaving")
18:43:29preglowmy music habits aren't so predictable that i can just put down a three hour playlist and know i'm not going to want to listen to something else halfway
18:43:41 Quit san|chez (Read error: 110 (Connection timed out))
18:44:19n1spreglow: that's why we can insert and remove tracks in aplaylist :)
18:44:34n1sinsert space where appropriate
18:45:38preglowsure, but i can't be bothered to insert, i just play it
18:45:40LloreanOne can not use playlists, and at the same time, not use the auto-change-directory
18:46:18LloreanI'd still kinda like a one-click "Play this folder" (possibly optional) for targets that have both "Right" and "Select" as enter folder.
18:47:21mokkurkalvetried to install rockbox on e280 - can't boot rockbox nor original firmware with |<< Guess I'll have to use recovery mode. Used sansapatcher on Linux and it said "Bootloader installed successfully"
18:47:25preglowLlorean: i'd dig that too
18:47:40 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
18:47:42 Part pondlife ("Gone")
18:47:55Lloreanpreglow: There's currently a one-click-insert patch, but iiuc it doesn't work very well (breaks menus for people who use the insert button as their normal navigation button there, etc)
18:48:00preglowLlorean: think that'd be a nice action for "select", really
18:48:06preglowLlorean: it makes more sense for "right" to enter
18:48:11LloreanAnd I'd rather it be one-click "Start playing this now" rather than insert.
18:48:28preglowagreed, agreed
18:49:36pixelmamokkurkalve: "boot _with_ |<<" ? Maybe just a misunderstanding but you press power before and then hold the "left" button?
18:50:34n1sLlorean: personally I'd like one-click inser more that one-click-start-playing
18:50:40mokkurkalveyes. "Cant load /system/OF.bin"
18:51:09mokkurkalvewon't boot either firmware now
18:51:43Lloreann1s: There's already a very quick method of inserting files or folders anyway. A one-click insert just makes it easier to accidentally mess up your playlist and not realize you've done it yet.
18:52:08Lloreann1s: Whereas one-click play-now, you'd know it at least. Not to mention Party Mode would make it one-click insert instead of one-click play-now, I should hope
18:52:19n1sLlorean: so a one-click-start-playing would not mess up yout playlist?
18:52:37Lloreann1s: Read what I said. The key isn't whether it messes it up, but whether you know that you've done it.
18:52:58n1salso to just start playing something is currently a lot easier than inserting...
18:53:01mokkurkalveCan't load rockbox.m4i Cant load rockbox.e200 - when I try to start normally
18:53:04LloreanIf you accidentally insert, say, the next track in an audiobook, you could end up listening to the book out of order and not being aware at all.
18:53:13Lloreann1s: To just start playing a folder?
18:53:30Lloreann1s: The whole point of this was having Select just start playing a folder, while Right enters it, rather than both acting the same when you click on a folder.
18:53:33n1sjes, click select, click select on first track, voila
18:53:53Lloreann1s: And "Artist" folders with six albums in them, that I want to play the whole artist? Or whole genre?
18:53:58LloreanYou're making assumptions...
18:54:17mokkurkalveI have ofcourse extracted .rockbox to the root before running sansapatcher
18:54:32LloreanBesides, if "Select" is *always* Play Now, and "Right" enters folders, "Right" could also insert when done on an individual song, perhaps
18:54:39n1sLlorean: that is just a one click insert with a "kill your playlist first"
18:54:58LloreanStop playback, then insert.
18:55:00*n1s thinks that is just wierd
18:55:14LloreanSo... clicking on a song to start playing it isn't weird, but clicking on a folder to do so is?
18:55:42n1sso whit a regular one click insert you could have your thing with a simple press of stop first
18:56:05n1sLlorean: i think so, yes
18:56:27Lloreann1s: And with existing, a long press and two short presses inserts anyway. Which is already very fast.
18:57:34 Join DerPapst [0] (
18:58:20LloreanI think a lot of users just play music, rather than fiddling with playlists...
18:58:33n1sLlorean: as it is basically the only thing i do in the file browser I think it is slow enough to be annoying
18:58:52mokkurkalveguess I'll have to follow "recovery mode" instructions? I downloaded SKU_E-PP5022.mi4, this I should copy to recovey patition?
18:59:25Lloreann1s: Considering many MP3 players don't even have on-player playlist building, I think the more "normal" method is just to play songs that the user wants to play, than to spend time inserting, etc.
18:59:34n1sLlorean: I'm just talking about what I think, i have noi ide what the users do...
18:59:37 Join mf0102 [0] (n=michi@
19:00:08LloreanAh, well I'm thinking at least that a "Play Now" would be more useful, to most users, than an "Insert this" one-button. But that's of course still an opinion.
19:00:28LloreanWe get enough people asking how to play all their music, or a whole artist, I think just telling them 'Click select on the artist folder' makes a lot of sense.
19:00:34 Quit Frazz (Read error: 110 (Connection timed out))
19:00:34LloreanEspecially if having shuffle on will shuffle it, etc.
19:01:01advcomp2019mokkurkalve, you just need to start over like you said
19:01:12hcssounds like something that would be a suitable configurable option, "select does... a) insert into playlist b) play directory c) w/e"
19:01:28n1sTalking about the users I think most would expect a SELECT click on a dir woul open it not play it...
19:01:51Lloreann1s: I think left and right to navigate in and out of folders is acceptable.
19:02:00LloreanWith select actually "invoking" whatever you click it on
19:02:09n1shcs: yes, I think this would be one place where a configurable button would be nice
19:02:34Lloreann1s: Again, why not use party mode then?
19:02:39 Quit J3TC- (Connection timed out)
19:02:50hcsLlorean: ah, I always forget there's that mode
19:02:51n1sLlorean: because I like being able to skip tracks.
19:03:17 Join Psy-Krow [0] (
19:03:25LloreanOkay, see, that's so unlike my listening style I can't describe it.
19:03:42LloreanIf I shuffle a large folder or something, yeah, I skip around. But if I actually craft a playlist, I'm very unlikely to want to skip
19:04:19hcsThe rest of us are bad predictors of our own preferences.
19:04:24n1sI usually insert a lot of alnums into a playlist and then turn on shuffle...
19:04:27PaulJamLlorean: but why do we need a special button to do something that is already possible with playlist=>insert in the context menu? it'll save only one (or two) buttonpress.
19:04:47LloreanPaulJam: I'm NOT suggesting a special button to insert...
19:04:54LloreanPaulJam: That's n1s suggestion
19:05:24PaulJamLlorean: well, insert is the same as play when playback is stopped.
19:05:28LloreanPaulJam: I'm suggesting a special button to play the highlighted folder. Which saved either three shorts and a long, or two shorts and two longs depending on target.
19:05:38LloreanWHEN playback is stopped.
19:05:51LloreanBut the discussion is over whether it should insert or start anew when playback is already active.
19:06:16LloreanPaulJam: And, for that matter, can you defend why we should have two buttons that do the exact same thing as opposed to giving one of them a shortcut to a frequently desired feature?
19:06:34 Quit Psy-Krow (Remote closed the connection)
19:07:49 Join Psy-Krow [0] (
19:08:06Lloreann1s: Do you frequently insert songs, or folders?
19:08:11mokkurkalveWell... Now I got "Load main image fail."
19:08:21n1sLlorean: usually folders
19:08:47mokkurkalveCan't have been correct firmware file....
19:09:10advcomp2019mokkurkalve, do you have a R or plain e200
19:09:10PaulJamLlorean: because there were always two buttons doing the same and many people get used to either using the one or the other button. i always use NAVI/SELECT to enter a dir, and i wouldn't like if that behaviour changed (but i guess this isn't a very strong argument, maybe i'll come up with something better).
19:09:43LloreanPaulJam: The "Record" button always did nothing, and we changed that... it's hardly an argument at all
19:09:52pixelmaLlorean: nothing for the Ondio then (but you already said that) :)
19:10:07 Join Redbreva_ [0] (
19:10:15 Quit barrywardell (Remote closed the connection)
19:10:18Lloreanpixelma: I did in fact cover my basis by saying "targets where there are both a right and select button that do the same thing" :)
19:10:26LloreanThat ondio is tricky
19:10:38LloreanI need to get one just so I can use it for a while and stop forgetting about it.
19:10:51mokkurkalvenothing indicates an R. backside says "e200" and there was no mention of "Rhapsody" in the orignal frimware
19:11:13PaulJamLlorean: thats different, the REC button did _nothing_, so there was no reason to press it before.
19:11:41LloreanPaulJam: Rec could be held down to prevent USB connection.
19:11:43pixelmaPaulJam: I heard quite a few people used it to turn on the backlight :>
19:11:53LloreanAnd plenty of people used a short press for the backlight.
19:12:04 Quit hcs ("Leaving.")
19:12:14PaulJampixelma: a short press still doen't do anything.
19:12:17LloreanI think that any keymap should really be considered flexible up until a "release" version is made.
19:12:38mokkurkalveI did upgrade to latest official firmware before trying to install rockbox
19:12:38LloreanBeing afraid to change it just ties you into what may have been bad design decisions earlier on
19:12:42LloreanSuch as having redundant buttons.
19:13:05advcomp2019mokkurkalve, what version is that
19:14:03***Saving seen data "./dancer.seen"
19:15:16mokkurkalveI think the Sansa Firmware Updater said 01.02.18F was installed
19:15:51PaulJamLlorean: i just don't see the need. on h300 you save one long buttonpress if playback is stopped and two buttonpresses (short,long) if something is playing. I mean wasn't this your arguement yesterday when someone asked for a button to play all songs?
19:16:32LloreanPaulJam: You save A long and a short, or a long and two shorts
19:16:58LloreanOn iPod you save as much as two longs and a short.
19:17:06PaulJamjeah, i was wrong, you save one buttonpress more in bothe cases
19:17:31LloreanPaulJam: My argument yesterday was more pertaining to the fact that the button couldn't be used for anything BUT playing all songs with his recommendation
19:17:56 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:18:05LloreanBeing able to tread albums, or genres, or whatever as simply and cleanly as treating songs when you click them in the file browser or database does increase convenience
19:18:08Lloreantreat rather
19:18:46LloreanAnd this function would in fact solve his problem more elegantly by far
19:19:20n1sLlorean: using the button as a regular insert would be more flexible as you would only have to press stop first to get what you want, thus two short presses ore one long one short, depending on target
19:19:34LloreanPaulJam: And frankly, I don't see the need for preserving a redundant button, and that surely has as much wait as your not seeing the need for streamlining the ability to play things
19:19:58Lloreann1s: Honestly, I'd accept that if not for two things
19:20:03PaulJamwell, i think i'll just get used to it if someone implements it. i don't think i have any influence on the decision anyway :)
19:20:04mokkurkalveI can now only enter recovery mode, but it's no help if I cant transfer a working .mi4 file
19:20:05Llorean1) Stop is frustrating on iPods
19:20:30 Quit Domonoky (Read error: 104 (Connection reset by peer))
19:20:36Llorean2) There's discussion of preserving the playlist when playback is stopped, and enough in favour of it that it may happen, meaning that stop then insert would actually insert while keeping music stopped, and not clear the playlist
19:21:07Lloreann1s: On iPods it's entirely possible to accidentally shutdown while trying to stop playback.
19:22:49n1sLlorean: why stop at all if you are preserving the playlist? currently we start playing when something is inserted...
19:23:04 Join DM| [0] (
19:23:25Lloreann1s: Don't ask me, I'm strongly opposed to preserving the playlist when stopped
19:23:33LloreanI believe the argument is "to allow playlist editing while playback is stopped"
19:23:37n1sso one press on that button would just insert with an unecessary stop...
19:23:47n1sLlorean: we have pause...
19:24:00Lloreann1s: Again, don't argue to me, I'm not in favour of it.
19:24:13LloreanBut you can't pause *before* playback starts.
19:24:20n1sbut then there's the argument that pause == stop because both makes it go silent :>
19:25:11LloreanAnyway, even ignoring whether or not the Pause/Stop Clear/Don't Clear functionality changes, there's still the issue of the fact that on iPods long-play is Stop and longer-play is Shutdown
19:25:36LloreanIn fact, I'd wager the majority of iPod users don't know that "Pause" isn't stop, and that there IS a stop.
19:25:39n1sLlorean: move shutdown to another button?
19:25:52LloreanThere's not a better button
19:26:08LloreanStop is used less than Context Menu, FF, RW, or probably even Quickmenu
19:26:14LloreanAnd those are all the buttons you got
19:27:03PaulJammaybe something like doubleclick for shutdown would work?
19:27:07mokkurkalveYes. What I need is 01.02.18F which is "Europe (with FM)"
19:27:55n1sLlorean: IMHO the low number of buttons shouldn't make other targets' keymaps less usable...
19:28:11n1s*on ipods
19:28:39Lloreann1s: That's a faulty argument
19:29:02LloreanIt's shaping the added usability, rather than making it less usable.
19:29:28 Quit linuxstb (Read error: 110 (Connection timed out))
19:30:02n1sLlorean: you stated that the reason you wanted this button was because stop is annoying, it isn't on h300 for example so the more flexible solution would make both the stop-insert and regular insert easier
19:30:07LloreanConsidering on all targets it's equally easy to Insert, while on some targets it's harder than others to Stop before inserting...
19:30:26Lloreann1s: No, I didn't say the reason I wanted it. I said the reason I favourted it over simply inserting.
19:30:48LloreanYou're basically saying we should favour a specific player rather than picking the solution that works best all around...
19:31:54LloreanThere are no players with difficulty inserting, afaik, so making it less difficult is less "gain" than making playing a new folder (which already takes more button presses on all targets) easier.
19:32:10LloreanAnd one-click-playing is something people expect out of the UI anyway, as can be seen by frequent questions or expectation by people
19:32:11n1sLlorean: no, the reason I think it's a strange ide to inser one action that does something that I don't ever do myself and really don't think most people do either so it isn't "best all around"
19:32:18LloreanThough usually they expect to play something with the "Play" button
19:32:37Lloreann1s: What basis makes you think most people don't want to play folders?
19:33:28 Join J3TC- [0] (
19:33:29Lloreann1s: Have you not seen the new users who frequently ask how to play a folder, artist, or whatever, and have to be told to stop, then insert it? Or how many ask why pressing "Play" doesn't play the folder they have highlighted and have to be told it's resuming playback, and doesn't actually play?
19:33:50n1sLlorean: I think people like to add stuff to a playlist easy, this thing basically moves away from playlists to a "directory mode" sort of thing IMO
19:33:57 Nick fxb__ is now known as fxb (
19:34:02Lloreann1s: I asked "on what basis"
19:34:26LloreanThis means "what have you seen that suggests this". I've seen many users *expect* the functionality I described to already be there, and be disappointed that it's not.
19:34:40LloreanPersonally *I* have more use for "one-click insert" and would prefer it on a purely personal basis.
19:34:51n1sLlorean: I have made no scientific studies if that's what you are asking
19:34:54LloreanThe reasons I favour one-click play are based more on what I've seen apparently desired and needed
19:35:14LloreanI'm not asking for scientific studies. "People keep coming in here and expecting to be able to use it" is hardly a scientific study in favour of my side
19:35:25LloreanBut so far all you've told me is that *you* prefer it, and because of that you think that's how most other people would do it too
19:35:42 Quit Febs (Remote closed the connection)
19:38:09mokkurkalveOK. I can still enter "recovery mode" but the .mi4 image downloaded fails. What do do?
19:38:13 Quit japc (Remote closed the connection)
19:38:20PaulJamLlorean: btw, what makes you think that people DO frequently want to play folders that contain several subfolders with music? i havent seen many requests of that kind.
19:38:33LloreanPaulJam: As I said, I have seen many requests of that kind
19:38:52n1si have seen people using their players in very different ways but have never seen anyone use rockbox. I think most people accept the limitations and abilities of their respective player and so would expect to be able to do something similar to what they are used to. BUT when they have more freedom, like using computers ALL the people I have been secretly studying make playlists
19:38:56LloreanPaulJam: Including people who expect to be able to, and are surprised when the can't, and people confused because pressing "Play" on a folder resumes playback of something else rather than playing that folder
19:39:18Nico_PLear: hi, I see you closed FS #8158... I guess things are good for you too now?
19:39:49advcomp2019mokkurkalve, do you have the right pp5022.mi4 file
19:40:19LearNico_P: Yes, can't reproduce it any more. No more MoB-related problems for me. :)
19:40:24 Join styleism [0] (
19:40:25Nico_Pcool :)
19:40:32PaulJamLlorean: i hope you don't want to change the function of the play button.
19:40:52LloreanPaulJam: I never said that I did.
19:41:35PaulJamLlorean: good, your last anwer somehow sounded like that.
19:41:50mokkurkalveI have the one downloaded from page linked from rockbox wiki page. It's obviously not the right
19:42:05LloreanMy point was just that from all the time I've spent helping new users, I frequently encounter confusion that there's not some button to "Play" a folder.
19:42:26LloreanBecause in a *lot* of OFs there's a "Navigate" button that shows you the list of songs (or albums) by an artist, and a "Play" button that just plays that artist
19:42:56 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
19:43:49mokkurkalveWhat I had before running sansapatcher was firmware 01.02.18F, which is for european e200 with FM. Don't know how to get that .mi4 file
19:44:09LloreanAnd while duplicating OF functionality is not necessarily a strong reason to do something, I see a lot more people expect to be able to do that, than expect to be able to one-click insert.
19:44:41LloreanAnd I very honestly think that a one-click insert will produce a lot of confusion among new users. "Why, when I click on the folder, does nothing at all happen?"
19:44:48LloreanInsert is essentially transparent if music is already playing
19:44:59LloreanAnd it may be 100 songs before you get to it.
19:45:05 Join webguest02 [0] (i=18b2733e@gateway/web/cgi-irc/
19:45:23mokkurkalveWhat I downloaded now was SKU_E-PP5022.mi4, and it don't work...
19:46:00advcomp2019mokkurkalve, did you rename the mi4 file to pp5022.mi4
19:46:30mokkurkalveno, I didn't
19:46:48mokkurkalveI'll try that
19:46:48advcomp2019mokkurkalve, that might be your issue
19:47:57webguest02i have a sansa e280r with rockbox installed. things look fine (small but fine) however i can't locat or play flac files. they exist in the musc dir. where can i find reference to play flac on this device.
19:48:20n1sLlorean: We will never be able to make our UI not confuse users and we already expect them to read the manual so I don't think that is a strong point.
19:48:41Lloreann1s: But it's not a non-point either.
19:49:09advcomp2019webguest02, you need to show all files since the music folder is hidden
19:49:15n1sLlorean: I agree and I also disagree with you and I don't think this discussion is going any further :>
19:49:22Lloreann1s: A new user can at least pick up Rockbox right now, and from trial and error get most of the functionality pretty quickly. A button that provides no apparent action, and no user feedback adds confusion
19:49:28bertrikwebguest02: AFAIK you should be able to just select a flac file to play it, not any different than opening an mp3
19:49:46Lloreann1s: I can agree with that last point. It's good to have all the cards on the table.
19:49:56PaulJamLlorean: do you think it would be a good idea to start a poll which of the two buttons should do the play now action? i personally would prefer the right button, but i don't know how others think about it.
19:50:00webguest02when i use the database the flac files do not show up.
19:50:29webguest02i would reply to specific nicks but i don't recall how. been too long since i used irc.
19:51:12webguest02i thought i should just be able to play the files. must be something wrong on my end.
19:51:45 Quit stewball`ghost (Read error: 110 (Connection timed out))
19:51:49LloreanPaulJam: Left and Right should stay "navigation", in my opinion at least. I'd rather not poll it, because honestly the feedback you get isn't really useful
19:52:05bertrikwebguest02: are the flac files tagged somehow?
19:52:06LloreanPaulJam: When you poll, you're getting feedback from a subset of users that aren't really indicative of "all users" in any way.
19:52:26LloreanThe average user could have completely different expectations, but also doesn't expect to have to check the forums to make sure things don't change.
19:52:39webguest02yes. the flac files are tagged with artist, title, track and some other info.
19:52:50LloreanIt's kinda better just to make a decision based on whatever other factors you can include than any sort of "guess" at "user preference"
19:53:01PaulJamLlorean: but then you can say "Hey, there was a poll, so don't complain" :)
19:53:31LloreanPaulJam: Yeah, but at the same time, after you start the poll, you may find a significant reason to do it the way other than the way they voted.
19:53:42LloreanI think it's better to just try to come up with objective reasons for one way or the other
19:53:51bertrikwebguest02: did you add the flac files recently? perhaps the database is not updated yet
19:54:28webguest02added the flac files last night and have done two updates to the database but they don't show up.
19:54:42bertrikwebguest02: In that case, I'm out of ideas
19:54:48 Quit amiconn (Nick collision from services.)
19:54:49LloreanPaulJam: I received slightly more positive feedback than negative for my e200 keymap change. One week before I posted final warning to the mailing list and forums. People didn't really respond until *after* the change was in SVN
19:54:57 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
19:55:12pixelmawebguest02: also what type of tags are those (Rockbox only supports vorbis comments in flac files)?
19:55:16LloreanPaulJam: And then, when I said "You have months to respond" the response was invariably "How was I supposed to know?". Basically, keymaps will always step on toes.
19:56:26webguest02ah. perhaps that is it. i think the comments are the flac formatted ones. i'll create a set with the vorbis comments and see if that works. i have to access the comments using metaflac.
19:57:37pixelmawebguest02: and to see the files in the file browser, you need to either put them in a different folder (the sansa original firmware marks the "music" folder as hidden) - or as already suggested set the file view option to "all"
19:57:53pixelmavorbis comments are the native tag format for flac
19:59:07webguest02the flac files are in the music dir. i'll try changing the view option for files. thanks.
19:59:08 Join Bagder_ [0] (
19:59:55 Join przemhb [0] (
20:00:10 Quit J3TC- (Read error: 104 (Connection reset by peer))
20:00:40 Join J3TC- [0] (
20:00:43webguest02ok. that did it. once i set the file view option to all they showed up.
20:01:17przemhbamiconn: what do you think may cause FS #8238 ?
20:02:10amiconnUmm, no idea...
20:02:33*amiconn didn't try wake up alarm at all, on any target
20:02:57amiconnAnd I always have startup screen == browser
20:04:05webguest02thanks for the help. this will make using the device a lot simpler. now i won't have to convert from flac to mp3 before i listen to music. excellent! have a good day folks! thanks for the help.
20:04:21 Quit mokkurkalve ("leaving")
20:05:00przemhband have you got any idea concerning tuner's i2c problems some users experience?
20:05:06 Quit webguest02 ("CGI:IRC")
20:05:45preglowprzemhb: shouldn't really be a wakeup alarm thing, wakeup alarm just calls the screen like usual
20:05:52 Join Robin0800 [0] (
20:06:05przemhbfor example FS #8144 or FS #8177
20:06:44n1smaybe some init isn't done in correct order or given enough time when waking up
20:07:48przemhbit may be - normally tuner starts in some 1-2 seconds
20:08:10przemhbfirst you hear only noise and then a station tunes in
20:10:00przemhbmaybe a new i2c data is send to the tuner before it tunes-in and so spoils frequency setting?
20:10:35 Quit advcomp2019 ("Never look down on someone unless you're helping them up.")
20:15:37 Join miepchen^schlaf [0] (
20:15:52 Quit Bagder (Read error: 110 (Connection timed out))
20:15:59 Join linuxstb__ [0] (
20:17:40 Quit Robin0800 (" HydraIRC -> <- In tests, 0x09 out of 0x0A l33t h4x0rz prefer it :)")
20:18:03 Join Robin0800 [0] (
20:23:27 Quit Robin0800 (Read error: 104 (Connection reset by peer))
20:25:33 Quit BigBambi (Remote closed the connection)
20:25:41 Join TMM [0] (
20:26:43 Join toffe82 [0] (
20:27:51 Quit linuxstb_ (Read error: 113 (No route to host))
20:28:29 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:31:48 Quit J3TC- (Read error: 110 (Connection timed out))
20:31:53 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
20:32:28 Join J3TC- [0] (
20:36:16 Join JdGordon [0] (
20:36:43 Join lids_ [0] (n=lds@
20:36:43 Quit lids (Read error: 104 (Connection reset by peer))
20:38:13 Join Rincewind [0] (
20:40:46 Join Arathis2 [0] (
20:41:44 Join iamben_ [0] (
20:47:13 Quit J3TC- (Read error: 104 (Connection reset by peer))
20:47:38 Join J3TC- [0] (
20:48:11 Quit spiorf (Remote closed the connection)
20:49:28 Quit ompaul (Read error: 113 (No route to host))
20:51:01 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:53:11 Quit iamben (Read error: 110 (Connection timed out))
20:53:25amiconnprzemhb: Hmm, I really have no idea. Tuning, scanning and switching between presets works here both at 30MHz and at 80MHz
20:53:46 Join adrenergic [0] (i=sorry@
20:53:48amiconnAnd starting in FM screen returns to the last played station fine
20:54:11adrenergici need help
20:54:22przemhbso it is very dependend upon H10 unit...
20:54:40adrenergicanyone can hear me :$
20:54:56przemhbmy friend's H10 tuner hardly works
20:55:13 Join Robin0800 [0] (
20:55:28pixelmaadrenergic: just ask and be patient as per the guidelines (see topic)...
20:55:29adrenergicanyone that can help me :(
20:55:30amiconnMaybe there's a difference in hookup between the 5/6 GB and the 20GB
20:55:41amiconnI have a 6 GB
20:55:44adrenergicoh ok
20:55:51przemhbI don't think so
20:56:08przemhbI've got 5G - I'll try it
20:56:15pixelmaadrenergic: someone who knows will answer
20:56:22przemhband see how tuner works on it
20:56:27adrenergicwhere can i find the TAA english version for Gigabeat MEG200J
20:57:30 Quit Arathis (Nick collision from services.)
20:57:37 Nick Arathis2 is now known as Arathis (
20:58:02amiconnprzemhb: My build is r15832 if that matters. And btw, the i2c clock cannot be too fast
20:58:34pixelmaadrenergic: does that have anything to do with Rockbox? Besides, I have no idea what version you are talking about...
20:58:36amiconnWe use udelay(2) for each half-period, that makes (at least) 4 us for the full preriod, i.e. (max) 250kHz
20:59:06amiconnThe TEA5767 can handle fast i2c (up to 400kHz), and the driver handles clock stretching
20:59:15 Quit J3TC- (Read error: 104 (Connection reset by peer))
20:59:38przemhbamiconn: I have tried to increase it up to 50 - it did not helped my friend's H10 to work any better...
20:59:51amiconnI thought that possibly the SDA line might also miss a pullup on some H10's, but that can't be
20:59:57 Join J3TC- [0] (
21:00:11amiconnYeah, as I'd expect
21:00:19amiconnThere must be a different problem.
21:00:24amiconnIs that a 20GB model?
21:01:01amiconnOkay, so I *bet* there's a difference between the 5/6 and the 20
21:01:20amiconnjhMikeS, who commented on fs #8177, also has the 20GB model
21:02:06 Part adrenergic
21:03:02przemhbmaybe there were introduced some changes to H10 circuits during mass production(?)
21:04:31 Quit Seed (Nick collision from services.)
21:04:37 Join Seedy [0] (
21:05:09 Join FOAD_ [0] (
21:05:50 Join stripwax [0] (
21:06:20amiconnThere could also be a difference in rockbox
21:08:18przemhbthere was no RB version that has worked good for my friend, so the problem exist from the begining
21:11:06przemhbI have just checked how my second H10 (5GB) behaves - simply saying just fine
21:11:19przemhbthere's no problem at all
21:12:51 Join CrumbCat [0] (
21:13:20 Quit CrumbCat (Client Quit)
21:13:44 Join Zagor [0] (
21:14:05***Saving seen data "./dancer.seen"
21:14:25 Join Rob2222 [0] (
21:14:55amiconnpreglow: The button driver uses non-atomic GPIO handling. Not sure whether this might cause that kind of problems, but atomic port handling is better anyway. Will change and test
21:18:35 Join Febs [0] (n=chatzill@
21:21:50 Join obo [0] (n=obo@rockbox/developer/obo)
21:22:03 Quit FOAD (Read error: 110 (Connection timed out))
21:22:03 Nick FOAD_ is now known as FOAD (
21:22:27preglowamiconn: what problem?
21:22:44amiconnUnstable tuner operation on H10 20GB
21:23:20LloreanWas that maybe supposed to have been directed at przemhb?
21:23:51pixelmamy guess too :)
21:23:52amiconneh, yes
21:23:58 Quit mf0102 ("Verlassend")
21:24:01 Quit Zagor ("Client exiting")
21:24:08preglowamiconn: oh, right, that certainly is possible and worth doing anyway
21:24:09amiconnToo many nicks matching pr<tab>
21:24:36 Quit J3TC- (Read error: 110 (Connection timed out))
21:24:52preglowamiconn: don't worry, i am very sympathetic to that problem :P
21:24:56 Join desowin [0] (
21:24:56amiconnRight now I just did the H10 button driver, because it uses GPIO D like the fm tuner driver
21:25:29amiconnThere are a lot more places which could (and probably should) be converted
21:26:20 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
21:28:33 Quit Rob222241 (Read error: 113 (No route to host))
21:30:47 Quit zicho (Remote closed the connection)
21:31:17 Quit JdGordon ("Konversation terminated!")
21:31:42 Join zicho [0] (
21:32:55 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
21:34:14 Join stewball [0] (n=WTFOMGBB@
21:38:41 Join japc [0] (
21:43:05 Quit amiconn (Nick collision from services.)
21:43:13 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:44:33amiconnIs tuner operation stable on e200, or does it have the same problems as on c200?
21:44:48Lloreanamiconn: Unstable how? Slight drift?
21:44:59amiconnNo, flaky operation
21:45:26 Quit DM| ("*bashes head against keyboard*")
21:45:31amiconnSometimes working sometimes not, sometimes even requiring to enter the radio in the OF to get radio back in rockbox
21:46:00LloreanI've heard some people say that the signal is a lot weaker than it is in the OF. And others say that if they load a preset for 95.5 or some such, that it'll actually tune in a bit off, but inconsistently so that it's not a constant offset.
21:46:12LloreanAs well people have had their radio appear and disappear across boots.
21:52:53markunanimeloe: what are you trying to say?
21:58:27advcomp2019animeloe, my radio is off by .1 mhz
21:59:47advcomp2019oops, amiconn i mean
22:01:04 Join Zagor [0] (
22:03:26przemhbamiconn: I've checked what is H10 tuner's frequency drift on alarm wake-up - it must be in -1.45 -1.7MHz range
22:04:54przemhbthe actual frequency is 1.45 to 1.7 lower from the required frequency
22:06:30przemhbatomic GPIO setting did not improve anything
22:09:54jhMikeSamiconn: did you even try setting the delay in the lv24020lp driver to 2? the udelay of 1 may often end up too short. I asked low_light to check that but never got word if it was.
22:12:41amiconnWell the datasheet says that the 3wire bus can operate up to 700kHz
22:13:04amiconnudelay(1) should keep the clock <= 500kHz
22:13:30amiconnI can't try that now, because I normally don't have access to a Sansa
22:13:37amiconnpixelma could test that
22:14:13pixelmasure, what do I need to change and where?
22:14:21jhMikeSyeah, but udelay(1) will actually delay until the next uS which may be too short
22:15:44 Join J3TC- [0] (
22:16:39pixelmaok, will try in a bit
22:16:54 Join Matteo_M [0] (
22:17:24 Quit Matteo_M (Client Quit)
22:18:07 Join Matteo_M [0] (
22:18:12Matteo_Mhi guys
22:18:23*jhMikeS hopes its that simple anyway :)
22:18:41Matteo_MI've just downloaded rbutil for osx, and I have problems installing themes
22:19:09Matteo_Mit gives me a network error, but I have installed with no problem the rockbox firmware and fonts
22:19:16Matteo_Many suggestion?
22:20:33 Quit styleism (Read error: 110 (Connection timed out))
22:20:54amiconnjhMikeS: Any idea about the PP5002 cache problem yet?
22:20:58Matteo_Mhere a shot that can explain:
22:21:12amiconnI put an svn bootloader on my 2nd Gen - no change :(
22:21:29jhMikeSamiconn: that's turning out be a bugger on how to get it started without corrupting things. :\
22:22:30Matteo_MMaybe the problem is a bug of rbutil that can't connect to the right server in order to download themes...
22:22:57LloreanMatteo_M: I think the theme server is just going through changes right now
22:23:30Matteo_Mok, is what I was expecting to heat llorean, thank you
22:23:46 Quit Matteo_M ("Leaving")
22:25:56 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
22:29:54Nico_PjhMikeS: any clues as to why queue_wait causes a data abort and sleep freezes on the S?
22:31:21 Quit w0rd54 (Remote closed the connection)
22:31:29 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
22:33:16Nico_Plinuxstb_: ping
22:34:47 Nick Bagder_ is now known as Bagder (
22:34:48jhMikeSNico_P: on the S? can speculate it's something to do with core sleep
22:35:13Nico_PjhMikeS: yes, on the S. what does that mean?
22:35:45 Join sandisk [0] (
22:36:27 Join w0rd54 [0] (
22:36:27 Quit sandisk (Client Quit)
22:36:28jhMikeSno ports to new processors have bothered implement sleep_core
22:36:31pixelmajhMikeS, amiconn: increasing that delay doesn't help
22:36:50jhMikeSpixelma: argh...worth a short anyway
22:37:29Nico_PjhMikeS: ah I see now... is it really critical? if so, what should I look for to try implementing it?
22:37:36*jhMikeS can't see the point in not implementing the basics before doing everything else...yes it's boring as hell, but needed
22:37:54*amiconn agrees
22:38:21*Nico_P will try, but needs a few pointers
22:38:31pixelmaI can quite reliably reproduce if I listen to some music files before trying to enter the radio -> frequencies off. Maybe that helps someone to get an idea...
22:38:32Nico_Pand I need to learn ARM too :p
22:38:41jhMikeSNico_P: it's spelled out on how the implement it in thread.c
22:39:05 Quit Lear ("ChatZilla 0.9.79 [Firefox]")
22:39:34jhMikeSjust fill in the right part with the processors sleep command
22:39:53Nico_PjhMikeS: thread.c:411?
22:40:51jhMikeSthread.c 100. these have to be done according to how the processor behaves (whether or not it may sleep with interrupts disabled and such).
22:42:05Nico_Pah yes... I'll try doing it then
22:42:06jhMikeSreason for all that is to never have to wait out a tick when something should be running _now_ which is quite a long wait
22:42:28 Nick parafin is now known as parafin|away (
22:43:00 Quit linuxstb__ (Read error: 113 (No route to host))
22:43:09Nico_PjhMikeS: it's likeley to be like the other ARM versions, isn't it? and is the interrupt disabling/enabling code the same,
22:44:09 Quit stewball (Read error: 113 (No route to host))
22:45:57jhMikeSs2c2440 and pp are quite different in the approach
22:48:30jhMikeSthat one with not wake up if interrupts are disabled when the core clock is cut, but it takes several cycles to actually stop. that sets the idle bit then enables the interrupts in the next instruction. if one is pending it will basically abort the idle.
22:49:15jhMikeSSH is similar to that. PP and Coldfire can just leave them disabled.
22:51:20jhMikeSthe idea is to keep the interrupt pending at the time the core is trying to sleep because that interrupt may have made a task ready to run
22:53:22 Quit rasher ("leaving")
22:55:38 Join rasher [0] (n=rasher@rockbox/developer/rasher)
22:58:42 Quit nicktastic ("Leaving")
22:59:20 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
22:59:38*ender` yawns
23:00:15 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
23:00:34 Quit Redbreva_ (Read error: 104 (Connection reset by peer))
23:00:41 Quit advcomp2019 (Nick collision from services.)
23:00:49 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
23:05:17 Join stripwax [0] (
23:06:35 Quit Arathis ("Bye, bye")
23:06:42 Quit desowin ("use linux")
23:07:00 Join webguest90 [0] (i=47bf1606@gateway/web/cgi-irc/
23:07:58 Quit webguest90 (Client Quit)
23:08:32jhMikeSNico_P: any chance there's still a problem with the BSS inits on the S?
23:08:59Nico_PjhMikeS: I have no idea... I'm not even what the implications would be or how to check :/
23:09:16Nico_Padd "sure" after "even"
23:09:32Nico_PI'm a complete noob
23:10:08jhMikeStry enabling the kernel and thread checks and see if you get any interesting panics
23:10:37Nico_Pwhere are the checks?
23:11:04jhMikeSa #define at the top of kernel.c and thread.c
23:11:06 Quit linuxstb_ (Remote closed the connection)
23:11:38 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
23:13:29stripwaxDoes H1xx support albumart?
23:13:39Nico_PI get a data abort which I'm pretty sure comes from queue_wait
23:14:07***Saving seen data "./dancer.seen"
23:14:30Nico_Pstripwax: look in config-h100.h... you can also enable it there if it's not
23:14:50jhMikeSNico_P which queue wait?
23:15:08*Llorean can't imagine Album Art looking particularly good at 2bpp
23:15:22rasherLlorean: in some cases it's quite recognisable
23:15:25stripwaxIt seems to 'work'
23:15:36rasherIn others, not so much.
23:16:10stripwaxAny devices have a bitmap lcd but not albumart?
23:16:44 Quit male ("User disconnected")
23:17:10Nico_PjhMikeS: I get a data abort at 8804140C... the .map seems to place queue_wait there IIUC
23:17:31*stripwax guesses recorder
23:18:12stripwaxsliding_puzzle seems to build and run fine for h120, just testing for recorder. Anything else I need to take a look at?
23:18:17linuxstbstripwax: Just do "grep HAVE_ALBUMART firmware/export/config-*.h" to see which targets have it.
23:19:22 Quit jhMikeS (Nick collision from services.)
23:19:28 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:19:29 Quit Fraser (Read error: 110 (Connection timed out))
23:19:50stripwaxlinuxstb - right −− thanks
23:20:12 Quit idnar (Nick collision from services.)
23:20:14 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
23:22:00 Quit nicktastic ("Leaving")
23:22:09stripwaxGigabeat-S and mrobe500 don't HAVE_ALBUMART but do have 16-bit displays - is that expected?
23:22:32Nico_Pstripwax: I didn't enable it because it's not very useful ATM
23:22:42stripwaxtpj1022 also
23:22:48jhMikeSNico_P: of course it could be a bad pointer jumping it there too
23:23:19stripwaxNico_P - what isn't useful - the build or the albumart?
23:23:34Nico_Pstripwax: AA on these builds in their current state
23:24:09Nico_PjhMikeS: should system-arm.h:set_interrupt_status work ok on the S?
23:24:24 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
23:24:29n1sNico_P: are interrupts working at all?
23:24:37Nico_Pn1s: I don't know
23:24:58Nico_PI seem to recall aliask saying they didn't, but I don't know how to check
23:25:21jhMikeSNico_P: it should work on any ARM I'd think
23:25:30stripwaxNico_P - is that just because the build isn't working fully? should sliding_puzzle on these targets be able to access the cover.bmp files, or not? I think my patch won't build on those so I'll try that.
23:26:09n1sNico_P: i'm not too familiar with these things either but there should be some sort of interrupt status register we could check.
23:26:22Nico_Pstripwax: I don't think you need to care much about these targets right now... also if your code is enclosed in #ifdef HAVE_ALBUMART, it should be fine
23:26:26jhMikeSIt follows ARM's own guidelines re: the way to set the cpsr
23:27:32Nico_PjhMikeS: is there a way I can check interrupts work?
23:27:52stripwaxNico_P - right. sliding_puzzle currently does not test HAVE_ALBUMART for the albumart-specific puzzles and I wanted to see if HAVE_ALBUMART was defined always whenever bit depth > 1 (seems not though)
23:28:07stripwaxAdding HAVE_ALBUMART sounds like it will be necessary after all. thx
23:28:23jhMikeSon that you can put a splash screen in the handler for the interrupt your want to check
23:29:09Nico_Pstripwax: HAVE_ALBUMART is basically independant from any other #define. I added it in a bunch of config files and target maintainers can do it for recent ports. if you use AA you need those #ifdef checks
23:29:23 Quit bertrik ("bye")
23:29:30jhMikeSaren't ticks running?
23:29:52Nico_PjhMikeS: I don't know... sleep seems to be problematic though
23:30:02Nico_Pit seems t simply freeze the player
23:30:07stripwaxjust wanted to understand why it's a separate #define that's all
23:30:33jhMikeSNico_P: no panics then I take it?
23:30:48Nico_PjhMikeS: haven't seen any, I'll disable queue_wait to check though
23:31:04 Join zajacattack [0] (i=42617a92@gateway/web/cgi-irc/
23:31:36Nico_Pstripwax: it's separate to allow per-target fine-tuning
23:31:46stripwaxfair enough
23:31:54zajacattackzagor: what remains left to be done with the USB stack?
23:32:14jhMikeSIs the S allowing nested interrupts to occur (not disabling interrupts in interrupt mode)?
23:32:29Zagorzajacattack: get it to work
23:33:11Lloreanzajacattack: Please don't come in here every day and ask that. It'll be done when it's done.
23:33:18Nico_PjhMikeS: as per my previous experiments, no more data abort with queue_wait disabled, and no panics either
23:33:27 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
23:33:40zajacattackLlorean: i know; i only asked cuz yesterday zagor wasn't here
23:33:57Nico_PI have a few printf calls that allow me to say the player freezes in ata_init(), and I'm thinking it's on the sleep() call
23:34:19*jhMikeS wonders about that comment in kernel-imx31.c "the following needs to be fixed". It's completely archaic.
23:34:37Nico_PjhMikeS: regarding nested interrupts, I have no idea
23:35:15jhMikeSNico_P: any idea which queue_wait is causing it though? Is the queue one with queue_send enabled?
23:35:56Nico_PjhMikeS: it's hard to say, the panic error message clears the screen so I can't see which steps were done OK
23:36:03 Quit J3TC- (Read error: 110 (Connection timed out))
23:36:20*jhMikeS notices no interrupt ackknowlege code on the S
23:36:52jhMikeSwhich panic?
23:37:04Nico_Perr sorry I meant the data abort
23:37:27jhMikeShmmm...make it not clear it?
23:37:42 Quit hcs ("Leaving.")
23:37:49 Part zajacattack
23:37:59Nico_Pdo you know where that error message is defined?
23:38:09jhMikeSin system.c
23:38:49jhMikeSthe interrupt code is rather fishy looking
23:40:37 Join Lars_G [0] (n=Lars@unaffiliated/lars-g/x-000001)
23:40:45Lars_GI see lots of work on the usb side of things. how's it comming along?
23:41:21jhMikeSok, it's EPIT_HANDLER...hmmm
23:43:28 Join DM| [0] (
23:44:13 Join Skail [0] (
23:46:10jhMikeSthe avic code appears that it should be using set_interrupt_status during the avic calls or at least setting INTDISNUM _before_ removing a handler
23:47:35jhMikeSeh, blah, no...looks fine. (sucks trying to evaluate for a device you don't own) :p
23:49:11jhMikeSthough, do we really need dynamic handler installation?
23:49:46Nico_PjhMikeS: the data abort seems to happen because of ata_init too
23:50:22jhMikeSany part of ata_init in particular?
23:50:34Nico_Perr that I already knew... I'm starting to be sleepy
23:51:06*jhMikeS didn't know that though :)
23:51:30 Quit Rincewind ("bye")
23:51:39 Part Lars_G
23:52:06Nico_Pactually I didn't either :p I'm starting to get confused
23:52:43Nico_PI'll try to see more precisely where it happens
23:55:59*jhMikeS wonders if cache handling is right over there or a factor at all
23:57:52 Join kclaf [0] (i=kclaf@
23:58:59 Quit Skail ("Concoction recent by Pier.")

Previous day | Next day