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 2021-06-16

00:03:49 Quit akaWolf (Ping timeout: 268 seconds)
00:03:56 Join akaWolf [0] (
00:04:33 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
00:11:33 Join tchan [0] (
00:39:07 Quit uplime (Changing host)
00:39:08 Join uplime [0] (uplime@offtopia/nomad)
01:31:47***Saving seen data "./dancer.seen"
02:25:50 Quit advcomp2019 (Read error: Connection reset by peer)
02:32:56 Join j-r [0] (
02:37:04 Join advcomp2019 [0] (~advcomp20@user/advcomp2019)
03:31:50***Saving seen data "./dancer.seen"
04:05:31 Quit parity (Ping timeout: 272 seconds)
04:14:43 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
05:31:37 Quit kadoban (Quit: Bridge terminating on SIGTERM)
05:31:38 Quit blbro[m] (Quit: Bridge terminating on SIGTERM)
05:31:53***Saving seen data "./dancer.seen"
05:33:59 Join kadoban [0] (~kadoban@user/kadoban)
05:38:31 Join blbro[m] [0] (~blbrostra@2001:470:69fc:105::8f7)
06:04:28desowinI don't get where i got the 802.11b from, it looks like the chip is 802.11g
06:05:31braewoodsdesowin: really? that's more useful than expected.
06:06:05braewoodsit would mean WPA2 might be doable if we had an IP stack
06:10:58 Nick gevaerts_ is now known as gevaerts (~fg@user/gevaerts)
06:34:59 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
06:47:49desowinmurata docs I have say 802.11b/g and Marvell 88w8686 datasheet that's floating on internet says 802.11a/b/g
06:49:16desowinand it is on spi0 (toget with atmega) not on sdio
06:50:34braewoodsdesowin: so what do you plan to do with the wifi?
06:50:48braewoodswe would need a whole IP stack plus a wifi stack just to use it
06:51:48braewoodsthough it would open up some interesting options
06:52:41braewoodsgiven that i expect the main purpose of a driver is to optimize powersavings
06:53:11braewoods(if we can't use it, we may as well put it into its lowest power state we can muster)
07:19:04 Join cockroach [0] (~blattodea@user/cockroach)
07:27:49 Join dbohdan [0] (~dbohdan@user/dbohdan)
07:31:54***Saving seen data "./dancer.seen"
07:32:01dbohdanLooks like my Clip+ has finally died
07:32:16dbohdanIt served me well for a decade
07:32:33dbohdanThanks to everyone who has worked on Rockbox for making it so nice to use
07:32:57braewoodsdbohdan: so what can we do for you?
07:34:07dbohdanI just thought I'd show some appreciation for a piece for software I really like
07:35:41dbohdanI may have questions about what to buy to replace the Clip+ once I survey the local old audio player market
07:35:59braewoodsi see.
07:36:24braewoodsi have some HDD units that work well still. repaired and all typically.
07:36:45braewoodsif you live in the US i can sell 'em to you.
07:36:52braewoodslet me know what you find.
07:37:16dbohdanThanks, but unfortunately am cut off from international mail
07:37:26dbohdanThings would be much easier if I wasn't
07:37:35braewoodsah so you live elsewhere
07:37:35dbohdan(I am not in the US.)
07:37:45braewoodsok, some recommendations then
07:38:10braewoodsiRiver H100 series or H300 series are both good ports; old and reliable. main drawbacks are...
07:38:24braewoodsno softwate usb stack, only hardware UMS or MSC
07:38:53braewoodsbut they use standard connector types for power and data
07:39:04braewoodsno special cables needed
07:39:11desowinDoes someone speak Chinese here? seems to be available on some Chinese websites, but I have no clue how to get it...
07:39:40braewoodsdesowin: unfortunately no. we ran into a similar problem with a datasheet for a mips chip.
07:40:19desowinthere's also a newer one
07:40:23braewoodsdbohdan: the iriver H10 is also viable. stick to the fat model though, the 20GB one.
07:40:43dbohdanbraewoods: Noted. I'll check if there are any for sale on the local Craigslist clone.
07:40:52braewoodsdbohdan: those 3 irivers are the best option from iriver we have
07:41:27braewoodsthey have compatibility with the same battery type
07:41:42braewoodsand you can also CF mod them all
07:42:33braewoodsi would avoid any flash based units that lack SD card slots
07:42:45braewoodsthey can't be upgraded
07:43:08braewoodsall the sansa ports now though support multiboot for running entirely off the sd card
07:43:14braewoodsso you can preserve the internal flash
07:43:55braewoodsHDD units can be CF modded in most cases
07:44:06braewoodsSD card to CF card is also a thing
07:44:20braewoodsif real CF cards don't provide enough storage
07:45:09braewoodsoh, right. if you get the H100 or H300, make sure to flash the new bootloader before trying to do a CF mod.
07:45:36braewoodsi released new bootloaders that fixed some old bugs with it
07:45:40braewoodsand the OF doesn't like CF cards
07:45:46braewoodsso don't expect it to work with them
07:49:04dbohdanOkay, I've checked, and there are no iriver H10/H100/H300 players for sale near me
07:49:46dbohdanI think I'll just go through the entire list of supported hardware later and search for each model
07:55:14cockroachdbohdan: where do you live? I own an H140 that has seen better days but might be good enough for a bit of testing
07:56:18cockroachor, I could also test things myself, if you tell me what needs testing
07:58:31dbohdancockroach: I live in the LPR
08:00:38cockroachoh wow, that looks like an interesting place
08:01:26dbohdanIt is
08:01:51dbohdanSadly, we can't simply buy things off Amazon or eBay and have them delivered
08:05:14cockroachyeah, I was just wondering about that. international shipping must be a pain.
08:19:23sporkfiio m3k might still be available
08:20:21dbohdancockroach: Yes, it typically involves an intermediary
08:20:54dbohdanWe have one that specializes in AliExpress packages, so that's an option, but I haven't used them
08:23:06dbohdanspork: I can only find local ads for the FiiO X1
08:24:46 Join speachy [0] (~speachy@
08:24:46Mode"#rockbox +v speachy" by ChanServ (
08:25:40dbohdan(If you meant that for me.)
08:28:32speachysince you're the resident gui expert. :)
08:29:20braewoodsdbohdan: ask here first though. some models are less desirable for various reasons.
08:30:02sporkdbohdan: local is always hard, m3k might be on aliexpress still if that is an option
08:32:33Bilgusspeachy thats an oldie but a goodie I think
08:33:17BilgusIIRC I had to do special ordering with viewports to work around that bug
08:33:43Bilgus(in themes)
08:34:59dbohdanbraewoods: I definitely will!
08:36:32dbohdanspork: I agree about local. I still want to try local first, though, because it is faster and you don't have to worry about a package getting lost.
08:38:37speachyBilgus: One has to wonder how many more of these longstanding "features" are still lurking, eh?
08:39:02BilgusIDK how many features do we have?
08:39:18speachydunno, all of them?
08:39:18Bilgusmultiply by 100 and you might be close
08:40:51Bilgusthat theme engine is an unforgiving SOB too
08:43:45speachyjust for giggles, a first pass with sloccount says we have a bit under 1.5M lines of code
08:44:25speachy~1.4M C, 57K C++, 31K asm, 8K perl, 7K sh, and some miscellaney.
08:49:18BilgusI'm thinking scroll_stop should be called unconditionally
08:49:58speachyI've seen plenty of places where scrolling ends up drawing over a fade, for example
08:54:56Bilgusno clue why github couldn't bring me to this in search
08:55:23Bilgusits like their search function is worse than ever
09:01:08BilgusWe couldn’t find any code matching 'scroll_stop' in Rockbox/rockbox ?
09:01:42BilgusWe couldn’t find any code matching 'printf' in Rockbox/rockbox
09:01:56Bilgussomething mis configured?
09:03:49 Join vitt13 [0] (~vitt13@
09:07:55vitt13'Does someone speak Chinese here? seems to be available on some Chinese websites, but I have no clue how to get it.' I have account on , post me the link to download what you need
09:11:51Bilgusthere is the issue I believe
09:12:49 Join massiveH [0] (
09:18:17speachyie we stop for the current viewport but leave the rest scrollnig?
09:18:28mendel_munkisBilgus: I stopped using github code search years ago due to lack of results
09:18:46Bilgussame but it seems even worse
09:19:01BilgusI'm not on my machine either so no local source
09:19:11mendel_munkiswell all the people who want good results stopped using it...
09:19:17mendel_munkisoh ouch
09:19:42Bilgusspeacht clear_viewport is supposed to keep track of all intersecting viewports
09:20:31mendel_munkisprintf site: ?
09:20:39mendel_munkison google
09:21:13Bilgusyeah thats how i did it lol
09:23:12Bilgussorry that was condensed a bit too far clear_viewport uis supposed to call scroll_stop_rect which should pick tup the colliding viewports and stop them
09:26:26Bilgusif (s->vp == vp
09:26:46Bilgusthat seems suspicious
09:27:46speachyso that can only ever stop the current vp
09:28:24speachyeven if it's in the rectangle
09:28:33speachyyeah, that does seem wrong
09:30:58desowinvitt13: sent you link in private message
09:31:32speachyso... if you take away the vp check there, that'll effectively make the call into scroll_stop_rect()
09:31:58***No seen item changed, no save performed.
09:32:30speachyand scroll_stop_viewport becomes effectively the same as a global scroll stop
09:32:37Bilgusgood point
09:33:05Bilgusso perhaps its more nuanced than that though
09:33:56speachyI mean, there's probably a clear need for a non-vp-specfic scroll_stop_rect() but the question is if we actually need non-bounded vp-specific scroll stop
09:35:08Bilguswell I can see the need to only stop foreground elements versus bg but we currently have no notion of priority
09:35:55speachyso maybe instead of if (s->vp == vp) it could be (!vp || vp == s->vp) && ,,,
09:36:23speachyand we can implement a global scroll_stop_rect() that way. if it's actually useful, anyway
09:36:54speachythere aren't many places where scroll_stop (or its variants) are called anyway
09:37:47BilgusI think that will work without breaking anything
09:38:54Bilgusyou don't really want to stop everything just stuff that will overwrite your vp
09:38:57speachybut as you said the problem might ultimately be that there's no priority. stop_scroll_viewport_rect() stops the entire scroll line, not just the portion that's within the bounding box
09:39:39speachycould that result in scollers that never resume?
09:40:27BilgusI hate to say it but really what should be done is to keep track in the scroll engine and have it implement priority
09:40:55Bilgusyou place a new vp in the scroll area it wont scroll into it till it is removed
09:42:59speachyhow does it end end back up in the scroll list after the vp is removed? won't it have to be re-added?
09:43:42BilgusI was just debating that
09:44:25Bilguscurently we just let you make viewports as you see fit we would need to keep a list
09:45:05Bilgusif we turn vp into an opaque container and give a generator and destructor
09:46:20Bilgusthen we leave them in the list and they won;'t redraw if there is something blocking them
09:48:27Bilgusthat also fits well with the vp rewrite
09:48:52speachyhmm. or maybe.. just add a "scroll inhibit" flag to the scroller, and make the vps themselves check/clear the scroll list upon creation/destruction time?
09:49:10speachycheck/clear the inhibit in the scroll list I mean
09:50:00Bilguscouldn't you make it automatic though just check the scroller on vp creation
09:50:12speachythat's what I meant
09:50:55speachyI mean, the "Correct" solution is proper vp prioritization so scrollers can do whatever they want
09:51:13 Join petur [0] (
09:51:35Bilguswe don't have a real idea when a vp is destroyed though do we?
09:51:52Bilguslike how would we un-inhibit?
09:53:25speachyhow do vps get registered with the display code?
09:53:32speachysurely there is a de-register too?
09:53:45BilgusI guess on vp change you would clear the flag
09:54:00Bilgusno you just select the next one no deregister
09:54:28Bilgusclear the flag and check again for intersections
09:54:36speachyI mean, if the vp isn't "active" then who cares? :D
09:54:37Bilgusit'd still be glitchy
09:54:42Bilgusbut better
09:55:04Bilgusat least the active one would be clear
09:55:06speachyhmm. would this just be another layered hack
09:55:26Bilgusit all is thats why I said 'I hate to say this'
09:55:37speachythat we (meaning you) would have to undo when inevitably implement priorities?
09:56:20Bilgusbut its gonna take me 6 months to get to it
09:57:35speachybut going back to scroll_stop_viewport_rect() for a moment, you changed the semantics of that fucntion, dropping the "all lines if y < 0" clause
09:57:48speachyare any callers still expecting the old semantics?
09:58:03speachyoh wait nm.
09:58:06speachyline -> rect
09:58:21Bilgusyeah not my commit
10:01:07Bilgusultimately they could just prioritize their viewports as I did to work around the bug originally but what if we let them specify priotities in the theme engine
10:01:15Bilgusprobably a bigger can of worms
10:04:22speachywe'd need it independently of themes though?
10:04:51speachyand, heh, I doubt most themes are going to ever get updated.
10:05:47Bilgustrue unfortunately
10:36:53 Quit massiveH (Quit: Leaving)
11:05:46 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
11:12:28 Join skipwich [0] (~skipwich@user/skipwich)
11:27:32 Quit skipwich (Quit: DISCONNECT)
11:27:59 Join skipwich [0] (~skipwich@user/skipwich)
11:31:59***Saving seen data "./dancer.seen"
11:39:23 Join markun [0] (
11:49:22 Quit markun (Quit: leaving)
11:49:31 Join markun [0] (
12:27:24 Join reductum [0] (
12:39:47 Join advcomp2019_ [0] (~advcomp20@user/advcomp2019)
12:42:42 Quit advcomp2019 (Ping timeout: 240 seconds)
12:45:07 Quit JdGordon (Quit: Client closed)
12:47:43 Join F3l1x_10m_ [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
12:51:06 Quit F3l1x_10m (Ping timeout: 240 seconds)
13:32:00***Saving seen data "./dancer.seen"
14:47:32desowinapparently, the wifi module uses sif0 (15 MHz) while AVR uses sif1 (200 kHz)
14:48:22desowinand well, this wifi is like the perfect use case for ROCKCOM I mentioned earlier...
14:49:52desowinanother use case on Sansa Connect would be e.g. sending raw commands to AVR
14:51:09 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
15:08:59desowinbraewoods: Marvell firmware supports WPA2, so there's like nothing to be done on the host side, except from providing the password
15:09:41desowinbasically to get network going the Marvell driver has to be ported, hopefully the source code is GPL licensed so no problem here
15:10:06desowincombine that with lwip and you have networking, add mbedtls and you have tls
15:10:49desowinhowever, I might go only for the driver at most, leaving the rest for someone else :-)
15:11:53desowinthe firmware permits binary redistribution so that's probably ok for Rockbox (I guess that's first thing that would need binary blob)
15:13:36desowinthat's just three files in .rockbox directory (LICENCE.Marvell, gspi8686_v9.bin, gspi8686_v9_helper.bin)
15:16:14desowinofcourse that needs a ton of work on the user interface, which obviously need volunteer
15:32:04***Saving seen data "./dancer.seen"
15:36:32 Join amachronic [0] (~amachroni@user/amachronic)
15:46:06 Join lebellium [0] (~lebellium@2a01:cb10:2e:2000:5592:283e:3f8e:ad1c)
15:48:41amachronicanybody have objections to g#3474? I'd like merge it and g#3415
15:48:44rb-bluebotGerrit review #3474 at : Enable float formatting in printf by Aidan MacDonald
15:48:44rb-bluebotGerrit review #3415 at : Implement x^y in calculator by Moshe Piekarski
15:49:16speachydo it!
15:49:57rb-bluebotBuild Server message: New build round started. Revision a3f2b64a46, 297 builds, 9 clients.
16:01:56rb-bluebotBuild Server message: Build round completed after 719 seconds.
16:02:01rb-bluebotBuild Server message: Revision a3f2b64a46 result: All green
16:02:14rb-bluebotBuild Server message: New build round started. Revision 9ccae0421a, 297 builds, 9 clients.
16:11:51rb-bluebotBuild Server message: Build round completed after 578 seconds.
16:11:53rb-bluebotBuild Server message: Revision 9ccae0421a result: 0 errors 19 warnings
16:24:03rb-bluebotBuild Server message: New build round started. Revision 7aade87863, 297 builds, 9 clients.
16:33:47rb-bluebotBuild Server message: Build round completed after 583 seconds.
16:33:49rb-bluebotBuild Server message: Revision 7aade87863 result: All green
16:42:06sporkamachronic: i tried usbboot on the q1 but all i got was this "no partition found. insert usb cable and fix it"
16:42:22amachronicinsert SD card with rockbox zip on it :)
16:44:08sporkdid that, after compiling bootloader for the spl file and normal for rockbox.bin, i did a make zip and extracted the zip to the sd card
16:44:15sporkmaybe it is too big ?
16:45:17amachronicmaybe your SD card has a wrong partition table or filesystem then
16:45:27amachronicor maybe it's a power up issue
16:46:01amachronicnot too sure what I left in gerrit I had to tweak the power up delays a bit to make it work when booting from flash
16:48:01sporki will try again with the card from my m3k
16:59:59rb-bluebotBuild Server message: New build round started. Revision fa9ddd5af9, 297 builds, 9 clients.
17:01:46sporkwith the m3k card it runs
17:02:39amachronicprobably a GPT partition table or exFAT filesystem on your other card then
17:03:23sporkexfat i think
17:06:23speachyhad to revert the header parsing cleanup
17:10:02amachronichmm what did it break?
17:11:05speachyIt didn't fix the original problem and regressed other files. by the time I fixed the regressions.. the original code was cleaner to follow
17:11:18rb-bluebotBuild Server message: Build round completed after 679 seconds.
17:11:20rb-bluebotBuild Server message: Revision fa9ddd5af9 result: All green
17:11:59sporkno sound, but it is fun clicking around
17:12:48sporkmaking the theme imoroved it
17:20:17speachyok, found the root cause.
17:20:42speachythe actual mp3 data starts after ~600K of headers, we bail at 128K.
17:32:07***No seen item changed, no save performed.
17:32:43mendel_munkis600K of headers!
17:34:23speachyID3v1 tag duplicated over and over and over I think
17:37:33 Quit lebellium (Quit: Leaving)
17:38:46speachy g#3486
17:38:49rb-bluebotGerrit review #3486 at : FS #13299: Handle MP3 files with more than 128K of headers before audio data by Solomon Peachy
17:40:57rb-bluebotBuild Server message: New build round started. Revision 2ed4bf8a43, 297 builds, 9 clients.
17:49:31 Quit amachronic (Quit: amachronic)
17:52:16rb-bluebotBuild Server message: Build round completed after 679 seconds.
17:52:18rb-bluebotBuild Server message: Revision 2ed4bf8a43 result: All green
18:19:44 Quit vitt13 (Ping timeout: 252 seconds)
18:43:30 Quit pablocastellanos (Ping timeout: 264 seconds)
18:48:13 Join pablocastellanos [0] (~pidgin@user/pablocastellanos)
19:31:29 Quit petur (Quit: Leaving)
19:32:10***Saving seen data "./dancer.seen"
21:00:12 Join tomato1 [0] (~tomato@user/tomato)
21:01:58 Quit tomato (Ping timeout: 272 seconds)
21:01:58 Nick tomato1 is now known as tomato (~tomato@user/tomato)
21:32:13***Saving seen data "./dancer.seen"
21:46:30 Quit cockroach (Quit: leaving)
23:20:43 Quit F3l1x_10m_ (Ping timeout: 272 seconds)
23:24:01 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
23:32:15***Saving seen data "./dancer.seen"

Previous day | Next day