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

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

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

#rockbox log for 2020-11-13

00:00:16 Join Rower [0] (
00:00:42 Quit [7] (Ping timeout: 260 seconds)
00:00:58 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
00:06:49***Saving seen data "./dancer.seen"
00:07:54 Quit Rower (Ping timeout: 272 seconds)
00:10:43braewoods_bilgus: checked it. white screen, no signs of life.
00:10:53braewoods_bilgus: i waited a bit.
00:11:04braewoods_bilgus: not sure what it was supposed to do.
00:11:41braewoods_bilgus: basically, on a successful build the LCD does these things.
00:11:47braewoodsit flashes (it seems to default to white BG)
00:11:52braewoodsthen it changes to black
00:11:58braewoodsthen it starts to print text
00:12:00braewoodson the LCD
00:12:06braewoodsand eventually it switches control to rockbox
00:13:01braewoodsso it starts as black on white and then flickers, changing to white on black
00:13:04braewoodsand prints text
00:13:06braewoodsthen boots
00:13:19braewoodswhy it's not working for BL is beyond me
00:13:27braewoodsbut it does work fine in the plugin iriver_flash
00:13:37braewoodsi see it working when you boot up properly
00:14:04braewoodswe need to fix this somehow though or else i can't do what i planned to
00:14:12braewoodssince it needs a functional LCD during boot to work
00:14:31braewoods(the H120 bootloader has a navigation menu, otherwise silent boot would be fine)
00:19:09braewoods_bilgus: i think we may want to check the code for BOOTLOADER exclusions or exclusives
00:19:22braewoods_bilgus: we may need to pull something into the bootloader for it to function correctly
00:19:27braewoodsor exclude something
00:20:01braewoodsthe fact the main firmware works suggests either some critical code is run in a different order
00:20:07braewoodsor it isn't done at all
00:20:09braewoodsor both
00:27:17braewoodsnothing stands out in crt0.S
00:27:24braewoodsit's probably not responsible
00:29:09braewoods_bilgus: one thing i noticed. the BL for the H1x0 and H3x0 both have manual power control. their power_init excludes the automatic drive power on that is present in the normal code.
00:29:27braewoodsi wonder if we could use the HD spinup to signal code has been reached
00:29:49braewoodsit may trigger whenever it is powered up
00:30:18braewoodsthe sound of that would be a viable way to send a simple signal
00:32:10braewoodsbut the fact i don't even heard that means the BL fails to reach the storage activation code later on
00:32:46braewoodsso yea... somehow the init code for the BL and/or LCD appears to be getting stuck
00:33:14braewoodsit doesn't even boot at all if i wait a bit
00:33:40braewoodseither it's doing something that causes it to behave erratically like a null PTR dereference
00:34:18braewoodsthere's an idea
00:34:46braewoods_bilgus: one thing i notice, the update functions only do something when the LCD is turned on
00:34:55braewoodsi wonder what would happen if the functions were no-ops for the BL
00:34:59braewoodswould it still boot?
00:35:22braewoodsbbl, got some stuff to do
00:39:16 Quit advcomp2019_ (Ping timeout: 256 seconds)
00:39:25 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
00:52:07 Quit atsampson (Ping timeout: 272 seconds)
02:03:10 Join Rower [0] (
02:06:51***Saving seen data "./dancer.seen"
02:30:48 Quit advcomp2019 (Ping timeout: 258 seconds)
02:31:48 Quit Topy44 (Ping timeout: 260 seconds)
02:42:46 Join Topy44 [0] (
02:47:15 Join ac_laptop [0] (~ac_laptop@
03:06:26 Quit ac_laptop (Ping timeout: 256 seconds)
04:06:54***Saving seen data "./dancer.seen"
05:06:54 Quit Oksana (Read error: Connection reset by peer)
05:23:16 Join edhelas [0] (
05:38:25 Join prof_wolfff [0] (
05:42:16 Quit Stanley00 ()
06:06:58***Saving seen data "./dancer.seen"
06:12:42 Quit mendel_munkis (Ping timeout: 272 seconds)
06:13:10 Join mendelmunkis [0] (
06:17:38_bilgusbraewoods, I figured it was an issue of the dma still but the math looks fine AFAICT
06:17:56braewoods_bilgus: i see
06:21:30_bilguswell another thing to try is to give it a different screen buffer static and see if it works
06:22:49braewoods_bilgus: do whatever you want to try. and if it would help you along i can send you one of the units.
06:27:04braewoods_bilgus: maybe DMA is just a red herring. i'm not sure at this point.
06:31:38 Join MrZeus [0] (~MrZeus@2a02:c7f:70d0:6a00:e8a4:66e6:8ec1:5d59)
06:32:55 Quit mendelmunkis (Remote host closed the connection)
06:33:39 Join mendelmunkis [0] (
06:34:54_bilgusI mean trying to not have you do that but maybe it'll make it a bit faster
06:35:34braewoods_bilgus: it won't cost me much. did you also want to test the LCD remote?
06:36:13braewoodshow about this
06:36:19braewoodswe see how far we get over the weekend
06:36:27*braewoods shrugs.
06:36:33_bilgussounds good
06:36:54braewoodsthough it was interesting how some changes to the LCD code in 2009 broke it for 10 years
06:37:02braewoodsand then again recently
06:37:08braewoodsso what really broke it?
06:37:30braewoodswe may not need the workaround if we can find out what really is going on.
06:39:01braewoods_bilgus: let me know when you got a new build
06:41:05_bilgushere is a build to try before I leave for the morning this one gives it a real 'slightly' oversized framebuffer
06:41:32_bilgusI can't actually decipher what line sized chunks actually means for the DMA
06:42:20_bilgusits entirely possible its falling off the end of the world and has always been doing that it just doesn't matter till the wrong thing is after it
06:46:49braewoods_bilgus: that was interesting.
06:47:06braewoodsit booted.. but the screen was random garbage until RB booted
06:47:24braewoodsthat's the main change
06:49:11_bilgusyeah there is just un initialized data there
06:49:26_bilgusof so entirely possible its a null pointer issue
06:50:03_bilgusnext one lets try making routing it back to the real framebuffer
06:50:11_bilgusbut bigger
06:51:12braewoods_bilgus: one thing to note, address 0 is mapped to the ROM on these targets
06:51:28braewoodsso reads to test will be returning data from the ROM
06:51:46braewoodswrites will silently fail if nothing else
06:52:04braewoodsyou need a really long command sequence to specific addresses to even modify it
06:52:07braewoodsso unlikely to harm it
06:52:36_bilgusI just came to a realization that the lcd code needs initd properly before HBADDR is valid I bet its missing
06:53:06braewoodsperhaps so
06:53:08_bilgusits the macro that gets the buffer
06:53:26braewoodsthat was the only direct change to the LCD code
06:53:33braewoodsthat i saw for H300 specific code
06:58:38_bilgushmm not finding the lcd_init
07:00:45_bilgusok another try
07:03:07braewoods_bilgus: i thought lcd_init was called from iriver_h300.c
07:03:50_bilgusit is and it is after the backlight init
07:04:37_bilgussame link
07:04:43_bilgusI think this should work
07:05:25braewoodswill try it. moment.
07:05:57_bilgusI think someone rewrote it to take into account when screen is off and neglected to make sure it was init'd before calling lcd code
07:08:19 Join atsampson [0] (
07:08:30braewoods_bilgus: it appears to working normally now.
07:09:07braewoodsthe lcd screen turns black
07:09:12braewoodsand prints stuff
07:09:16braewoodsthen loads rockbox
07:09:28braewoodsi wonder if we can try disabling my workaround.
07:09:32braewoodsfor DMA
07:18:14_bilgusI think your workaround is still needed
07:18:37braewoodswhat was your fix?
07:18:55_bilgusI'll rework this tonight and should have one more for you saturday morning
07:19:15_bilgusI moved the lcd_init to before the backlight init
07:20:01_bilgusI saw the patch yesterday about making lcd_off = backlight_off
07:20:50_bilgusdidn't occur to me they started pulling in the lcd stuff before init with that patch
07:23:58_bilgusup on gerrit bbl
07:37:33 Join ac_laptop [0] (~ac_laptop@
07:47:37 Join massiveH [0] (
08:06:59***Saving seen data "./dancer.seen"
08:20:42braewoodswow. the latitude 3350 is incrediblely repair friendly.
08:20:59braewoodsalmost everything is exposed for repair by just removing the bottom cover
08:37:18 Quit massiveH (Quit: Leaving)
08:43:17 Quit edhelas (Remote host closed the connection)
09:17:56 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
09:48:43 Nick mendelmunkis is now known as mendel_munkis (
09:49:54speachyok, v8 of my usb insertion prompt is up, now with sane bootloader handling and manual updates.
09:50:30speachyI'm tempted to change the default to mass storage (ie the existing behavior) and commit it as-is, even with the WPS wonkiness.
09:50:38braewoodsspeachy: meaning?
09:50:51braewoodsit will be selectable what it does when connected to usb?
09:52:04braewoodsjust keep in mind not all RB ports can charge over usb, if that's somehow relevant to what you're doing.
09:52:23braewoodse.g., H120 only charges over the barrel jack.
09:52:39speachyit's only enabled for targets with HAVE_USB_POWER
09:52:57braewoodsh300 should support that
09:53:23braewoodsin other news it seems bilgus found the bug making the H300 bootloaders fail
09:54:13braewoodsLCD is his area so i'll let him figure out how he wants to finalize the fix.
09:55:07braewoodsonce that's all said and done I can finally start updating the H300 BL
09:55:31braewoodsand work on pushing out a new set of bootloaders
09:55:33speachythe actual fix is pretty simple, but I don't know if he intends to commit the other changes he made along the way.
09:55:41braewoodsi'll wait and see
09:56:42braewoodsthe OF is still useful for one thing i've found
09:56:53braewoodsit's good for recovering from a bad bootloader build
09:57:02speachy...for flashing new images? :D
09:57:09braewoodsyea, basically.
09:57:24mendel_munkisjust don't force ask
09:57:26braewoodsbut once i'm done i'll be discarding the OF
09:57:36braewoodsthe OF is useful for development but not much else
09:57:53braewoodsi found a way to make the risk of bootloader bricking nearly 0
09:58:02braewoodsmade this whole debugging possible
09:58:35braewoodsi basically had to reset the unit quickly and power on again
09:58:40braewoodsit tricked the BL into booting the OF
10:00:06speachymendel_munkis: the patch currently defaults to ask (which IMO is more user-friendly) but there's a config setting to change that.
10:03:39braewoodsspeachy: so basically it sounds like you're implementing an android feature.
10:03:54braewoodsit won't expose file access over usb until you do so
10:07:00***Saving seen data "./dancer.seen"
10:31:21 Quit amsomniac (Remote host closed the connection)
10:50:40 Quit Huntereb (Read error: Connection reset by peer)
11:25:48_bilgusI'm going to pull the lcd stuff out of backlight_init and put it where it is supposed to be
11:30:16braewoods_bilgus: go ahead and i'll retest my own patch to see if it is still needed.
11:31:12_bilgusIts built on top of your patch you might just want to hit head and move the lcd_init and remote_lcd_int above the backlight
11:32:06braewoods_bilgus: it sounds like you were wanting to fix it in the subroutines and not the bootloader?
11:33:09_bilgusits going to be the same issue for bot it just happens that by the time the fw is running the lcd has been initialized by someone else
11:33:39_bilguswe shall make sure that doesn't happen again −−
11:36:24mendel_munkislooking at docs/CREDITS line 401 anyone know if that is one name or two?
11:38:42braewoods mendel_munkis: one. that's how it is presented in the commmits.
11:38:54braewoods5047b2e180ee72d869d6a50eabac5ebe5fdbd9cc ...
11:46:27 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
11:47:01 Quit advcomp2019 (Ping timeout: 264 seconds)
11:53:03_bilguswell it appears everyone else that has a screen that is not viewable when BL off does it the same so I will fix the other things I saw and just make sure backlight_init is after lcd_init
11:57:30braewoods_bilgus: incidently, there's some units that work w/o the backlight being on but not many
11:57:44braewoodsthe H120's screen doesn't need to use a backlight all the time
11:58:04braewoodsthen again it's a rather simple screen
11:58:09braewoodsmonochrome or grayscale
12:07:01***Saving seen data "./dancer.seen"
12:34:01braewoodsspeachy: regarding your USB changes... i'd go for 3 configurable options.
12:34:12braewoodsspeachy: Charge Only, UMS, and Ask
12:34:38braewoodsmeaning you can make it automatically choose for you or ask every time
12:43:07mendel_munkisthat's what the patch does
12:46:37braewoodsgo figure. pretty much all the MTP libraries i can find are C++.
12:49:11_bilgusbraewoods, at you leisure can you test the bootloade at the same link?
12:49:21braewoodsok. let me get it setup.
12:49:41_bilgusI found 4 other bootloaders with the same bug
12:52:43braewoods_bilgus: works.
12:52:53braewoodsnothing unusual though the print out seems a bit sparse
12:53:26_bilguslike something is missing now?
12:53:56braewoodsor maybe i'm just imagining it.
12:54:09braewoodsi thought maybe some printfs were still removed / commented out
12:54:14braewoodsthat's all
12:55:18_bilgusuh i'll check before I commit them real quick
12:56:03braewoods_bilgus: so in other words... my discovery of this bug fixed other bootloaders that would have been left unbootable as well?
12:56:49braewoodsjust i didn't originally know it would turn out to be a general issue
13:00:37_bilgusits hard to say but most likely
13:01:18_bilgusI didn't check if they all had lcd_update() in their backlight_init() but at least 2 did
13:01:47_bilgusso those would have surely crashed
13:02:47mendel_munkisthis is strange. in September I rebased my daily driver branch on master built the new toolchain and rebuilt. today I went to rebuild based on master and I see a bunch of what look like toolchain issues.
13:04:49braewoodsmendel_munkis: were you building for PP or ColdFire?
13:05:02braewoodsspeachy released a new toolchain in october
13:05:04braewoodsfor those
13:05:12_bilgusbraewoods, can you unset the WIP on your gerrit #2968
13:05:15fs-bluebotGerrit review #2968 at : h300: fix one long-standing bootloader bug by James Buren
13:05:29_bilgusits not letting me remove the WIP label
13:06:11mendel_munkisnope its arm
13:06:29braewoodsmendel_munkis: PP is arm
13:07:04braewoodsmendel_munkis: probably should delete the old toolchain and run the script again
13:07:11braewoodsit was changed in October
13:07:31braewoods_bilgus: how do you remove WIP?
13:07:36mendel_munkisit's not pp
13:07:42_bilgusset review
13:07:58mendel_munkisbut thanks for the advice
13:08:03fs-bluebotBuild Server message: New build round started. Revision f65fb2a, 293 builds, 8 clients.
13:08:10braewoodsmendel_munkis: uhm....
13:08:12_bilgusthank you
13:08:17braewoodsmendel_munkis: he updated the arm toolchain.
13:08:25braewoodsmendel_munkis: just PP is the most common arm target
13:08:29braewoodsthat i know of for RB
13:08:47mendel_munkisoh I though you where saying he updated the pp subset of the arm toolchain.
13:09:13braewoodsyou should try rebuilding it
13:09:17braewoodsit's based on GCC 4.9.x nw
13:10:01braewoodswe ended up disabling a new "feature" of GCC 4.9.x since it was generating bad code
13:10:03braewoodsfor now
13:10:22mendel_munkisyeah thats what I have. I just rechecked the calendar my last build was in october
13:10:31mendel_munkisbut I will rebuild and see what happens
13:10:48braewoodswell it was late october when it landed
13:11:10braewoodsaround october 13th
13:14:05_bilgusI'll be setting up new toolchains from scratch probably this weekend I'll try to do it as noob as possible (not that hard lol) and write down some notes
13:15:14braewoods_bilgus: did you check the H120?
13:15:25braewoodsit also has the backlight initialized before lcd
13:17:47_bilgusno I just greped backlight_init prior to lcd_init
13:17:54_bilgusbut I'll look
13:18:09braewoodsin current master
13:18:27braewoodsline 488 and 491
13:18:42_bilgusyep sure does
13:18:53_bilguslets see if I can up my grep skills
13:19:02braewoodsthough oddly the H120 actually worked with master
13:19:06braewoodsi dunno
13:19:18braewoodsmaybe just some targets have this issue
13:20:18braewoodsthough if i'm being honest some of these #ifdefs in the bootloader are pointless
13:20:25braewoodse.g., #ifdef CPU_COLDFIRE
13:20:26mendel_munkisso now I have to go check if my build server is up to date as well.
13:20:35_bilgusit all depends if someone optimized them
13:20:55braewoodsthe target for that BL is always a coldifre
13:21:03braewoodsso why is it useful to have that ifdef?
13:21:19_bilgustying lcd_enable(false/true) and backlight_on/off
13:21:38braewoods_bilgus: if you need me to I can test an H120 bootloader
13:21:48braewoodsi have a unit still with the OF so the recovery method should work for it
13:22:01_bilgusI'll just search a little deeper and submit another patch
13:22:20_bilgusbut feel free more testing the better
13:22:29braewoodsi'll need to anyway
13:22:32braewoodsat some point
13:22:40braewoodsit is pretty much feature complete
13:23:05braewoodsi plan to wait to build its final release for V7 until I'm done with the H300
13:23:15braewoodsthese will do a lot of the port
13:23:18fs-bluebotBuild Server message: Build round completed after 915 seconds.
13:23:22fs-bluebotBuild Server message: Revision f65fb2a result: All green
13:25:07 Quit bonfire (Quit: Leaving)
13:25:17braewoods_bilgus: serious question. i noticed you patched my fixes too.
13:25:33braewoods_bilgus: why'd you change these loops with an empty body with 2 ;;s?
13:25:49braewoodsthat's basically, empty loop statement / body and another empty statement outside it
13:26:05braewoodsthe second ; is technically not part of the loop
13:26:21braewoodsthough it does nothing so it will be optimized away
13:26:28braewoodsbut still, why? utterly redundant.
13:26:53 Join lebellium [0] (
13:27:07_bilgusit just helps show that it is an empty loop
13:27:14braewoodsI see.
13:27:39_bilgussome people do {} and others just a space
13:28:23braewoodsit's kinda funny how we have void / null statements just to satisfy the language grammar in some situations
13:28:43braewoodsshell has the same problem, you have a null statement (:) for these situations
13:29:17braewoodsotherwise there's no point to these
13:30:02_bilgusactually clang will throw warnings if a while loop is empty so its just something thats stuck
13:30:38_bilgusothers probably do too but I'm not familar with any
13:30:49braewoodsi just found shell a weird special case
13:31:11braewoodsit's a syntax error to have an empty loop / conditional / function body
13:31:25braewoodsthere's no way to denote an empty body like in C
13:31:39braewoodsyou have to provide a statement period
13:32:11_bilgusI'm sure if you go back far enough there was a very valid optimization reason when it was birthed
13:32:47_bilgusthat whole technical baggage thing
13:32:58braewoodsprobably to due with shell being line oriented
13:33:27braewoodsyou can use semicolons to put multiple statements on one line but that's the only time you'd use it
13:33:36braewoodsline endings normally terminate them
13:33:39_bilgushuzzah I moved to my text search and was able to find the others I missed
13:34:05_bilgusyeah I hate when people do that statement;statement;statement;
13:34:28braewoodsyou can also do this in C though i can't say i know very many situations where it's wise
13:34:33braewoodsthe comma operator
13:34:47braewoodsthe main time where you have to use it is in for loops
13:34:50_bilgusand I also find int x =0, y = 0; mildly triggering too
13:35:09braewoodswhy? it can be useful in that case. it's not a statement, it's a declaration.
13:35:31_bilgussure but just put them on their own line so you don't miss one
13:35:56braewoodsif anything i'd whine about how pointers work in this situation
13:36:10_bilgusIve also had situations of the compiler screwing up and making the rest an int after the first declare
13:36:10braewoodsunless it's part of a typedef the pointer does not propagate
13:36:38_bilgusI haven't seen that one in a long time but it left a lasting scar
13:36:52braewoodsyou mean the situations where C defaults to int?
13:37:04braewoodswhich are normally warnings these days
13:37:32braewoodsinteresting side note
13:37:46_bilgusno like this LONG y = 0, x = 0 and x becoming an int
13:38:00braewoodseh? i thought x would be long too.
13:38:14braewoodsinteger literals default to type int. you can only modify this by casting or adding an appropriate suffix.
13:38:31braewoodsL makes it become 'long', a second L makes it become long long.
13:38:38braewoodsand 'U' makes it unsigned
13:38:48_bilgusyeah and people often fail to do that
13:39:25_bilgushalf the issues in our lua port when I picked it up were from things like that
13:39:56braewoodsi can't say that seems to be an actual issue?
13:40:05braewoodslong x, y; produces 2 longs
13:40:08braewoodsfrom my tests
13:40:13braewoodsconfirmed with sizeof
13:40:14_bilgusI haven't run into that bug in a long while but I also don't tempt fate
13:40:22braewoodsoh. it's a compiler bug.
13:41:07braewoodsanother thing i learned about enums and integers
13:41:26_bilgusenums are always an integer type
13:41:36braewoodsthere's not much you can do to control the actual type the compiler chooses to use for your enum
13:41:38braewoodsi know
13:42:00braewoodsabout the only option i found was to assign a ridiculiously large value to it
13:42:04mendel_munkissame set of issues.
13:42:07braewoodsto increase the size
13:42:11_bilguswell you can always force it with a properly sized item
13:42:13braewoodsmendel_munkis: can you be more specific?
13:42:29mendel_munkisjust to make sure I should be running gcc 4.9.4?
13:42:41braewoodsi mainly work with coldfire
13:42:53_bilgusmendel what device is this?
13:43:18mendel_munkisfuze+. (the issues are with nonstandard plugins)
13:43:47_bilgusnon standard plugins?
13:44:02_bilguslol like ones you previously made?
13:44:03mendel_munkisrandom stuff from flyspray and suchlike
13:44:25mendel_munkisso this is all on me
13:44:29_bilgussimple question but did you try nuking the directory?
13:44:32braewoodsmeaning they're third party?
13:44:40*braewoods nukes it from orbit just to be sure.
13:45:20_bilgusthere are a few make clens that are very specific
13:45:43_bilgusso if you had objects from other builds lying around it wouldn't touch em
13:46:25_bilgusask me how I know :P
13:46:33mendel_munkisI did make clean but just nuked am trying again
13:48:29mendel_munkisyeah same stuff. pretty sure it's new warnings introduced in the new toolchain. I am just surprised I didn't see them before
13:52:25_bilgusodd just warnings though?
13:53:29mendel_munkiswell so far, besides for some stuff that's due to api changes
13:54:47 Join yang [0] (~yang@freenode/sponsor/fsf.member.yang)
13:59:20fs-bluebotBuild Server message: New build round started. Revision 47e1f96, 293 builds, 8 clients.
13:59:39_bilgusok that should be the rest of em thank you braewoods..
14:00:09braewoods_bilgus: did you use sed? i don't recall grep having any multiline options
14:00:36_bilgusno I used jedit
14:00:53_bilgusthen I was able to use (?ms)
14:01:02braewoodsi never knew people still used it
14:01:09_bilgusthats why I missed the other 4
14:01:35_bilgusI only use it for text searching I dislike its editor
14:01:47braewoodswhat do you use?
14:01:58_bilgusGeany ofc
14:02:05braewoodsi switched to sublime text a few years ago. best software i ever bought.
14:03:58_bilgusjust having a dark profile available helps a lot
14:04:17braewoodsfor geany?
14:04:24braewoodsST3 has plenty of dark color schemes
14:04:30braewoodsbut yea
14:04:32braewoodsit's not an IDE
14:04:57_bilgusfor whatever editor you prefer
14:07:05***Saving seen data "./dancer.seen"
14:13:31_bilgusspeachy you can go ahead and submit what you have and I can revisit it later to make it work with the wps but agree on the override setting I think it'd drive me nuts eventually otherwise
14:14:20fs-bluebotBuild Server message: Build round completed after 900 seconds.
14:14:22fs-bluebotBuild Server message: Revision 47e1f96 result: All green
14:14:58fs-bluebotBuild Server message: New build round started. Revision 6c3cc1c, 293 builds, 8 clients.
14:28:19fs-bluebotBuild Server message: Build round completed after 800 seconds.
14:28:24fs-bluebotBuild Server message: Revision 6c3cc1c result: All green
14:41:00_bilgushuh theSony A10 and A20 both have 3k size decreases from that skin patch any one with either device able to test HEAD?
14:43:35mendel_munkiswell the plugin author implemented rbendian.h on his own. and that wasn't even the issue
14:49:58yangI have an old MP3 player with FM radio, noname model "friend mp3 FD 100-B FM" which works on 1 x AAA battery. Would anyone like to adopt it ? I don't know how usefull would it be for your project...
14:51:13yangit has 128 MB RAM
14:51:58yangI am not aware if it can be reused in your project
14:56:28_bilgusyang no probably not especially if they aren't widely available thanks though
14:57:16braewoodslike i told them elsewhere... probably not. and would it even be practical?
14:57:21braewoodsthat won't hold much
14:57:32mendel_munkisok well I think I fixed that issue but now I'm seeing things I know I already fixed.
14:59:04yangok _bilgus
15:01:23lemon_jesushey gang, who's the resident expert on NAND drivers? I'm trying to RE the ipod nano 3g nand driver a bit and I've hit a wall where my knowledge runs out and google searches aren't turning up much.
15:09:52 Quit ac_laptop (Ping timeout: 260 seconds)
15:10:23 Join ac_laptop [0] (~ac_laptop@
15:11:33_bilguslemon_jesus, maybe wodz but I haven't seen our other one in a long while if one is lurking your best bet would be to idle a day or so and put a post on the forum
15:16:07lemon_jesuscool, I'll post to the forums and stick around. hopefully someone can help me out.
15:31:02bluebrotherhmm. So I'm looking into keeping voicefile installs separate in Rockbox Utilty. We do offer current builds and releases, but we have voice files for releases and daily builds. But don't offer daily builds for install. That's a bit ... inconsistent.
15:32:00bluebrotherso shouldn't we add daily builds back and offer prebuilt voice files for both release and daily builds?
15:35:56bluebrotherspeachy: can you add a [daily] section to build-info with the same items [bleeding] has?
15:36:06speachybluebrother: was about to propose something like that
15:37:17speachyso you'd want a timestamp and a revision that corresponds to whatever the daily build is? same format?
15:38:00bluebrotheryeah. So we can show that the same way we do for the dev builds.
15:39:13bluebrotherthinking about it, having something with the list of voice file languages might be good to.
15:39:43speachyI could generate several more languages than are being done, but I've only enabled those that I've been told sound okay
15:40:26bluebrotherif we have a list that are announced in build-info that wouldn't matter much :)
15:40:41bluebrotherand we could have that list in the [release] section too.
15:40:54*bluebrother checks things
15:42:25bluebrothercan you add voices=<english,english-us,francais,...>? Using the same strings we have in the voice file filenames? Both to [daily] and [release]?
15:42:37speachythere's already a build-info being generated, though not combined into the main build-info file
15:42:47speachydate = "20201113"
15:42:49speachyrev = 362f7a3
15:42:51speachyis that okay?
15:42:54bluebrotheroh. Right. I guess that was the file used before.
15:43:04bluebrotheryep, that's good.
15:43:20speachyokay, I'll update the daily script to combine that one in there too
15:43:49bluebrotherthen I'll add support for those.
15:44:06bluebrotherhave to rework the whole rbutil.ini first though, it's getting kinda messy :)
15:47:23 Quit livvy (Ping timeout: 240 seconds)
15:48:14 Join amsomniac [0] (~nadya@2601:181:8300:4db::b45)
15:48:15 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
15:48:25 Quit emacsomancer (Read error: Connection reset by peer)
15:49:53 Join emacsomancer [0] (
15:55:49speachybluebrother: build-info should have daily builds in it now
15:57:15speachyyou want the voices=whatever tag to be under release and dailies both?
15:57:56bluebrotheryes please.
15:58:45bluebrotheralso, can you change dailies/date to dailies/timestamp (so it uses the same name)? Would simplify things a bit for me.
16:02:10speachyis the abbreviated format okay?
16:03:52speachy(new one updated)
16:04:45 Quit t0mato (Ping timeout: 240 seconds)
16:05:17speachyfor the release voice, do you want it under the main [release] tag?
16:07:08***Saving seen data "./dancer.seen"
16:07:43speachyI think we may be better off putting voices in a separate hierarchy. voices/3.15=.... voices/3.14=... voices/daily=....
16:09:21 Quit emacsomancer (Read error: Connection reset by peer)
16:10:20 Join emacsomancer [0] (
16:10:52bluebrotherhmm. We only support installing the latest release, so having them separate wouldn't provide anything extra.
16:11:53bluebrotherbut then again, having them below [release] would kinda break things for retired targets. Unless we make sure we always have all announced voices for older releases as well.
16:11:54 Quit Rower (Ping timeout: 258 seconds)
16:12:08speachywell, "latest release" isn't the same for all targets
16:12:22bluebrotherthat's what got me thinking :)
16:12:36 Join t0mato [0] (t0mato@gateway/vpn/mullvad/t0mato)
16:12:50bluebrotherso a separate hierarchy would make sense.
16:13:11bluebrotherjust thinking how to get that done best in Rockbox Utility.
16:13:52speachyI mean, the voice list for all previous releases is just "english" but the next release will expand that considerably.
16:15:08 Quit emacsomancer (Read error: Connection reset by peer)
16:15:44 Join emacsomancer [0] (
16:17:23speachythe defined-but-disabled daily voice list consists of: deustch,norsk,russian,srpski
16:20:21bluebrotheryeps. And if someone is really motivated we could always add voice files for older releases :)
16:24:56 Quit t0mato (Ping timeout: 258 seconds)
16:27:04speachyokay, here's what's in the new build-info:
16:31:36 Join Rower [0] (
16:33:56bluebrothernice. Now I have to fix all the stuff I broke while restructuring :) Might take a while.
16:39:41 Join t0mato [0] (t0mato@gateway/vpn/mullvad/t0mato)
16:41:03fs-bluebotBuild Server message: New build round started. Revision fc4fff0, 293 builds, 8 clients.
16:47:12 Quit lebellium (Quit: Leaving)
16:52:35fs-bluebotBuild Server message: Build round completed after 692 seconds.
16:52:38fs-bluebotBuild Server message: Revision fc4fff0 result: All green
16:52:39fs-bluebotBuild Server message: New build round started. Revision 60f581e, 293 builds, 8 clients.
17:03:10 Join TheEaterOfSouls [0] (~souls@unaffiliated/theeaterofsouls)
17:11:49fs-bluebotBuild Server message: Build round completed after 1151 seconds.
17:11:58fs-bluebotBuild Server message: Revision 60f581e result: 12 errors 0 warnings
17:31:43 Quit livvy (Ping timeout: 240 seconds)
17:32:30 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
18:07:10***Saving seen data "./dancer.seen"
18:17:07fs-bluebotBuild Server message: New build round started. Revision 610ad6f, 293 builds, 9 clients.
18:24:25speachywhew, the only errors were due to a missing #include.
18:30:30fs-bluebotBuild Server message: Build round completed after 802 seconds.
18:30:36fs-bluebotBuild Server message: Revision 610ad6f result: 12 errors 2 warnings
18:34:14 Quit efqw (Quit: Connection closed for inactivity)
18:40:43_bilgusoh well it can onlt be all green for so long :p
18:45:13speachyyuck, looks like rewriting the ibasso usb code is needed to fix this properly. I think it predates the other hosted usb stuff.
18:46:34_bilgusI have all the aux menus for bluetooth done I still need to write the dynamic menus kinda stuck on the best way forward
18:47:56_bilgusI'm thinking I can use the multi-line select feature to display the name and mac but I need to figure how to make that co-exist with the single line stuff
18:48:55_bilgusspeaking of speach are you in a hurry for the rocker to be sent back/ on I think from here I can make due with the sim
18:49:04_bilgusspeachy *
18:49:07speachyno particular hurry
18:50:08_bilgusin that case I'll try to make a driver to go with it but that'll probably take me awhile
18:50:19_bilgusnot my forte
18:50:35speachythe ibasso stuff is wayyyy too expensive on ebay
18:50:56_bilgusprobably to do with rb TBH
18:51:15_bilgusmaybe you should ask for donated players on the frontpage
18:52:08speachy$115 is the cheapest dx50, dx90s are about $190.
18:52:25speachydon't think it's RB; more likely "audiophile" nonsense
18:52:46speachy(these were pretty high-end units in their day; modern iBassos are pretty expensive too)
18:52:59speachyyay, dinner's ready
19:03:21fs-bluebotBuild Server message: New build round started. Revision 43f9074, 293 builds, 9 clients.
19:26:39fs-bluebotBuild Server message: Build round completed after 1396 seconds.
19:26:46fs-bluebotBuild Server message: Revision 43f9074 result: 8 errors 2 warnings
19:30:03_bilgusoops sorry didn't mean to get in the middle of your thing
19:31:32 Part amsomniac
19:32:54speachysomehow ended up with fewer errors this time around
19:48:28 Quit prof_wolfff (Ping timeout: 256 seconds)
19:58:49 Join fs-bluebot_ [0] (
19:59:11 Quit bluebrother (Disconnected by services)
19:59:16 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
20:01:26 Quit fs-bluebot (Ping timeout: 265 seconds)
20:04:22fs-bluebot_Build Server message: New build round started. Revision 03cd773, 293 builds, 9 clients.
20:04:25speachyyuck, the dx50/90 port is definitely in need of some major TLC.
20:04:53speachyI think this should get it building again but the usb interaction needs to be rewritten in the model of the other hosted targets.
20:07:13***Saving seen data "./dancer.seen"
20:24:25fs-bluebot_Build Server message: Build round completed after 1203 seconds.
20:24:26fs-bluebot_Build Server message: Revision 03cd773 result: All green
20:29:54speachyso. wonder if it's worth using more rb funds to pick up a DX50.
20:30:54 Join S|h|a|w|n [0] (~shawn156@unaffiliated/shawn156)
20:41:27 Quit MrZeus (Ping timeout: 260 seconds)
20:46:11 Join massiveH [0] (
21:04:43 Quit ac_laptop (Ping timeout: 260 seconds)
21:52:04 Join captainkewl [0] (
22:07:16***Saving seen data "./dancer.seen"
22:08:22 Join ac_laptop [0] (~ac_laptop@
22:49:32 Quit captainkewl (Remote host closed the connection)
23:00:37 Quit ac_laptop (Ping timeout: 260 seconds)
23:41:43speachylovely, for some reason the server's http logs... didn't log anything for nearly a week
23:52:35_bilgustellin all of yall its sabotage!

Previous day | Next day