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 2025-02-18

00:31:10***Saving seen data "./dancer.seen"
00:56:29 Quit othello7 (Ping timeout: 260 seconds)
00:57:45 Join othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
02:00
02:31:13***Saving seen data "./dancer.seen"
03:00
03:44:14 Quit othello7 (Ping timeout: 272 seconds)
04:00
04:04:19 Quit jacobk (Ping timeout: 260 seconds)
04:05:01 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
04:15:39 Join OlsroFR [0] (~OlsroFR@user/OlsroFR)
04:16:32OlsroFRReceived today my new iPod (old) dock with the S-Video port and black big dot in front. This dock accepts both my modded USB cable and a real firewire brick ! Now even with screen constantly ON, it charges on Rockbox.
04:17:03OlsroFRThough with a real firewire brick, I get annoying electrical noises. But with the modded USB cable that charges at firewire speed, it's all fine.
04:20:01CH23interesting. I always chose the apple usb chargers because they were quiet. never tried with the firewire one
04:20:38OlsroFRI also received an USB charger, but for some reason, it does not work. The USB chargers also work great in Rockbox in many scenarios. Since I forced usb charging, it's perfect for all my older iPods.
04:20:52OlsroFRVideos, minis, 4th gen mono -> they can all charge reliably over USB on Rockbox :)
04:21:09OlsroFROnly the Classics are un-reliable
04:31:16***Saving seen data "./dancer.seen"
04:33:03 Quit OlsroFR (Quit: Client closed)
05:00
05:01:30 Quit olspookishmagus (Ping timeout: 246 seconds)
05:07:24 Join olspookishmagus [0] (~pookie@snf-137798.vm.okeanos.grnet.gr)
06:00
06:31:19***Saving seen data "./dancer.seen"
06:34:18 Join thanosengine [0] (~thanos@user/thanosengine)
06:39:43 Quit thanosengine (Quit: WeeChat 4.5.1)
06:55:24 Quit Moriar (Ping timeout: 245 seconds)
07:00
07:10:16 Quit PheralSparky (Read error: Connection reset by peer)
07:17:13 Join thanosengine [0] (~thanos@user/thanosengine)
07:26:33_bilgusCH23, that 'one' charger could be defective too
07:34:13CH23very true
07:44:34 Quit COMPL_EXE (Read error: Connection reset by peer)
07:44:53 Join COMPL_EXE [0] (~compl.exe@aosc/dev/origincode)
08:00
08:21:44 Join MarcAndersen [0] (~no_znepna@217.74.221.70)
08:21:55MarcAndersenAny
08:22:29MarcAndersenWhy is this happening with mingw? https://pastebin.com/H3w5VXCz
08:23:03MarcAndersenIt has something to do with the new things i bookmark.c or string-extra.h but I don't get it.
08:31:20***Saving seen data "./dancer.seen"
08:32:03 Join qf [0] (~qf@user/qf)
08:48:39dook_bilgus: thank you!
09:00
09:00:41_bilgusMarcAndersen, thats my fault, it doesn't match the std definition but apparently it still has a namespce conflict bc your system supploes it
09:01:01_bilgusI'll rename it and push a pach
09:01:14_bilguspaTch
09:01:20speachytbh I'm surprised it worked on SDL targets too
09:01:28speachywonder why windows is pulling stdlib.h in
09:03:33speachy_bilgus: imo should move our definition of itoa into our libc instead of the tagcache code.
09:04:16speachy(and move the implementation out of the bookmark code into libc too)
09:06:33_bilgusk
09:06:57_bilgusyou mean the fallback I assume
09:07:21_bilguscan't do that bc it gets used in plugins
09:07:52speachywell, it's a nonstandard variation so we can't fall back to a system itoa?
09:08:31speachy(...this is why I'm surprised it worked for hosted+sim builds)
09:09:57_bilgustruly they could just use printf like before its really on device where the extra speed and lower stack is noticeable
09:11:07speachyah I missed that the bookmark.c thing was a fallback, and the main implementtation was in a separate file.
09:11:07_bilgusI'll just not do it in libc at all and put it in misc with a different name, int_to_buf?
09:11:33speachyor rb_itoa()
09:12:03_bilgusjust that none of the itoas use a buffer like so
09:12:43speachyI just wonder about discoverability
09:13:08_bilgusgood point
09:13:28_bilguslseek is bad enough :p
09:20:31MarcAndersenSo that means that you had an idea about it would happen?
09:21:44speachy...honestly I'm surprised it worked for any non-native build given that itoa() is pulled in by stdlib.h normally
09:22:35MarcAndersenBut you think that my mingw is fine as it is?
09:22:58speachybut mingw has surprised us plenty of times with its own quirks. I still haven't updated my builders to the point where they can build a winsim successfully, so we don't have an automated way to catch crap like this.
09:23:17speachyyes, it's not a problem with your environment; the problem is quite clearly in the rockbox codebase.
09:23:26_bilgusI mean its got something pulling it in right but that shouldn't bother anything
09:24:15MarcAndersenI just wish I had the brain to learn c properly.
09:24:26_bilgusand for the record I had no idea it would I figured the non std definition would make it ok
09:24:51speachyC doesn't allow two functions with the same name.
09:25:03_bilgusyeah namespace conflict
09:25:05speachyC _macros_ however can have the same name if the number of arguments are different
09:25:30speachy...so if itoa() is defined as a macro, we're good.
09:26:50_bilguscompilers often provide it through the header <stdlib.h> while in non-conforming mode
09:27:00_bilguswhats non conforming mode?
09:27:52_bilgusmaybe thats the reason the difference in mingw or do we need that to even compile US
09:29:09speachywe are definitely non-conforming in bare-metal mode as we don't have a full c lib.
09:29:34speachy("non-conforming" with the C spec)
09:31:01speachythere is no itoa() in my system's headers btw
09:31:49speachy...so this looks like mingw defines it for $reasons.
09:32:39speachybtw, _bilgus, the font thing looks like a simple case of the two themes combined using more fonts than we have slots for.
09:34:45_bilgusI was figuring it just wasn't freeing them I haven't looked to see how its handling that currently but I think it will be only slightly involved mainly ensuring you don't free a font still in use
09:35:42speachythe bug report says "more than 5" but both using 6 seems like it ought to work
09:36:36speachyproblem is the "old" fonts aren't "unused" until the new theme finishes loading.
09:37:05speachymight be an off-by-one error in the accounting; haven't looked into it beyond the highest level.
09:37:22_bilgusyeah but I think the way its set up we can get away with it and it'll just be expensive
09:37:42_bilgusie its not in the cache so it reloads
09:37:44speachyhonestly, this seems like the answer is "bump the upper limit" or "don't do that"
09:38:01speachythe cache is independent
09:38:36_bilgusin theory as the theme is loading we can just not let anyone draw
09:38:39speachyeach font slot points at a cache
09:39:26_bilgusIDK I'll see once I look thru it wouldn't be the first time it was something totally different
09:39:31speachywe splash errors at some theme loading failures so we can't just stop the UI
09:39:46speachy(including font load failures, hah)
09:40:07_bilgusyeah but we can 'know' that we only have the inbuilt one by then
09:40:08speachyunless we effectively did the theme swap atomically.
09:40:32speachyare splashes aways using sysfont?
09:40:44speachy(if so, there's no point in translating most of 'em)
09:43:08speachygranted a theme load failure isn't really recoverable.
09:43:20speachy(at least not that I recall seeing...)
09:44:10speachyotoh, I only use cabbie on a 128x64x1bpp screen. :D
09:44:41_bilgus^ me too lol
09:45:15_bilgusthink mine is slightly squarer on the zip 96x96 maybe
09:45:51speachywell, that's at least color..
09:46:54MarcAndersenIs there a patch for the mingw issue or should I just wait to build? If I use a patch I think the builds would get an M.
09:47:54speachyhold off for now, bilgus is working on fixing it. we basically have to rename our itoa() function.
09:53:49_bilgusargh the attribution to groff means it needs to stay in its own fiel
09:53:55_bilgusfile*
10:00
10:28:20 Quit MarcAndersen (Read error: Connection reset by peer)
10:31:21***Saving seen data "./dancer.seen"
10:44:09_bilgus g#6389 look ok, its already non standard I just renamed it totally
10:44:12rb-bluebotGerrit review #6389 at https://gerrit.rockbox.org/r/c/rockbox/+/6389 : [BugFix] itoa conflicts with mingw rename to itoa_buf by William Wilgus
11:00
11:04:53speachylet's ship it. :D
11:11:23 Join Everything [0] (~Everythin@static.208.206.21.65.clients.your-server.de)
11:35:01_bilguswell almost got away with it too!
11:36:01_bilgusah I changed it to !plugin instead of !defined
11:38:37 Quit thanosengine (Ping timeout: 244 seconds)
11:40:31 Join thanosengine [0] (~thanos@user/thanosengine)
11:54:36 Quit thanosengine (Quit: WeeChat 4.5.1)
11:56:20 Join thanosengine [0] (~thanos@user/thanosengine)
12:00
12:02:30speachyinteresting, it resulted in a binary size bump
12:03:22hactar|antthere's no chance that rockbox will ever support combining characters is there?
12:03:51speachyin the sense of unicode composing you mean?
12:04:03hactar|antyeah. glyphs with negative x values, basically
12:04:18speachywe already do that for arabic and other languages with diacritic marks
12:04:28hactar|antoh really
12:04:29hactar|anthmmm
12:04:45 Quit thanosengine (Ping timeout: 244 seconds)
12:04:52hactar|antthe fonts system was impenetrable to me the last time i tried to read up on it
12:04:55speachybut full unicode normalization requires about a ~450K binary bump due to the necessary tables.
12:05:34hactar|antyeah i'm not looking for normalisation, in fact quite the opposite
12:06:29speachypoint being one needs to know _when_ to do said combining.
12:06:45 Join thanosengine [0] (~thanos@user/thanosengine)
12:07:36hactar|antis there rockbox-specific documentation on that or is it just part of the umm. i can't remember what the font format is.
12:07:58speachylook up is_diacritic() if you want to see how it's handled
12:08:17hactar|anttakk
12:09:10speachyarabic (and farsi, and some others) are unique in that the same "letter" can be rendered differently depending on where in the word it falls.
12:10:08speachyI don't recall if that's ultimately handled by transforming one codepoint to another or purely glyph substitution.
12:10:44hactar|antin my case i'm just looking to print a glyph in the x space of the previous, basically
12:10:48speachybut unicode normalization is essentially the same thing −− 1..N input codepoints -> 1 output glyph (which I _thikn_ is always a distinct codepoint as well)
12:11:26speachy(or 1 input codepoint -> 1..N output codepoints)
12:11:37speachy(depends on which normalization form you want)
12:13:29speachyok, diacritics are handled by "rewinding" the X position and overlaying.
12:14:39speachyand our implementation is not limited to Arabic.
12:14:57hactar|antand is that handled by the fnt file? or baked into the firmware
12:15:13speachybaked into the firmware. the font file has to contain all of the appropriate glyphs of course
12:15:34speachybasically the base letter and the diacritic mark have to have distinct unicode codepoints.
12:15:59speachyour fonts are 1:1 codepoint:bitmap
12:16:29hactar|antright, they don't contain metrics
12:16:55hactar|antsorry, metrics isn't the right word.
12:19:02speachybut where this ties into unicode normalization is that the same letter might be expressed as two separate code points (ie letter+diacritic)or a single codepoint that pre-combines the two.
12:19:09hactar|antright
12:19:25speachythey end up being rendered the same, but are different strings.
12:21:01speachyother than rendering diacritic marks (that table is likely way out of date fwiw) we don't have any awareness of unicode normalization. And it's just at the font rendering layer.
12:22:55hactar|antso if the combining characters in a font are mapped to codepoints that are not part of, say, arabic (or devanagari, etc), there's no way to use them as combining characters without implementing that in rockbox itself
12:22:56speachythis matters for things like filenames etc, where "รถ" might be U+00f6 _or_ u+006f + u+0308.
12:24:53speachyideally we'd take every string we get (eg from filenames, playlist contents, etc) and preprocess them to ensure they are all using a consistent normalization form. (eg "always combine" or "always decompose")
12:25:16speachythat way we can do meaningful string comparisons, sorting, etc.
12:25:35speachybut suffice it to say that's a lot of internal plumbing reworking.
12:26:10speachyI have a WIP patch that extends our unicode support to >16bit, currently insta-crashes on native builds. :D
12:27:03speachyafter that I want to layer on the utf8proc guts so we have the capability of doing this stuff, and then start work on the rest.
12:27:51speachybut we're looking at >75% jump in our binary size. A complete nonstarter on the lowest-end devices.
12:28:40hactar|antmaybe for 5.0 xD
12:28:45hactar|antoctober 2038
12:29:08speachy(one of the current hangups is dealing with the UCS2/UTF16 that FAT32 needs.
12:30:08speachy...our primary devices remain ~20-year-old designs.
12:30:24speachy(ie the ipods)
12:30:31speachybut at least they have RAM to spare.
12:31:16speachya lot of modern DAPs have firmware binaries in the ~40-60MB range; over half of that is a single unicode font.
12:31:22***Saving seen data "./dancer.seen"
12:35:21hactar|anti see dap projects these days and none of them seem to use rockbox which is silly when there's already an incredibly robust, free firmware out there. i guess the chips they're using don't lend themselves to supporting rb?
13:00
13:04:36 Join lebellium [0] (~lebellium@2a01cb0405d07f00a1d53b8ac411aea1.ipv6.abo.wanadoo.fr)
13:25:01 Join Burak__ [0] (~Burak@185.25.123.34)
13:28:02 Quit Burak_ (Ping timeout: 248 seconds)
13:37:42 Quit qf (Quit: Konversation terminated!)
13:41:54speachyoh, they're capable enoug
13:42:23speachybut the ones I've seen are pretty heavily tied to $microcontroller ecosystem, to the point of being completely non-portable
13:42:51speachyso getting a basic thing oging is simpler than porting rockbox
13:44:21speachyand it's always more fun to write your own something instead
13:44:31speachybut yeah, a lot of wheen re-inventing
14:00
14:31:24***Saving seen data "./dancer.seen"
15:00
15:11:39 Join MarcAndersen [0] (~no_znepna@217.74.221.70)
15:17:52 Join winsim-buildbot [0] (~lamest@85-218-172-116.norlyscustomer.net)
15:18:29MarcAndersenI got everything set up. It will do a build tonight.
15:27:53 Join akaWolf [0] (~akaWolf@akawolf.org)
15:34:39 Join othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
15:35:17 Quit Everything (Quit: leaving)
16:00
16:31:27***Saving seen data "./dancer.seen"
16:33:08 Join PheralSparky [0] (~S|h|a|w|n@user/shawn/x-4432647)
16:39:34 Join WebGuest58 [0] (~WebGuest5@128.127.104.138)
16:39:44WebGuest58hello guys
16:40:38CH23hello WebGuest58
16:43:16WebGuest58i have an ipod 6g with rockbox, recently updated the installed ancient build (it was from 2019) to one from february 2025, very nice indeed, thanks to all involved
16:45:49 Quit othello7 (Quit: othello7)
16:51:22 Quit MarcAndersen (Read error: Connection reset by peer)
16:55:08WebGuest58does anyone know which devices from here https://www.rockbox.org/daily.shtml have bluetooth? my only problem with my ipod is that i cant use my wireless earbuds
16:56:11XogiumI don't see any old players from the early 2000/2010s having bluetooth, tbh
16:56:27WebGuest58or if any of you owns a device that feels like it can be recommended to me
16:56:52WebGuest58Xogium: I see
16:57:16XogiumI could be wrong of course, but as rockbox primarily focuses on older hw
16:57:56WebGuest58the Shanling Q1 is on the list and from a quick glance it seems to have bluetooth, but i don't really like its form lol
17:00
17:02:49Xogiumright
17:06:08speachyrockbox has zero bluetooth support.
17:06:17speachyand that's not likely to change.
17:06:58speachyhttps://forums.rockbox.org/index.php/topic,55210.msg255427.html#msg255427 for the latest re-iteration of that.
17:07:44WebGuest58speachy: ooh, thanks for that, i wasn't aware
17:08:36Xogiumso if you want bluetooth, better stick to whatever's original firmware
17:10:49WebGuest58ok but what are some decent, solid brands? other than the ipod i know zero
17:11:41WebGuest58if you have any suggestions, if not it's okay
17:14:45 Quit kugel_ (Ping timeout: 246 seconds)
17:18:05 Join kugel_ [0] (~kugel@ip4d146a3a.dynamic.kabel-deutschland.de)
17:28:09XogiumWebGuest58: I mean, it really depends in what you would want in a player
17:31:43CH23still want to rebuild an ipod classic into an ipod shuffle
17:32:32Xogiumsounds like a fun project
17:32:48Xogiumbut why the shuffle, for the looks, or ?
17:33:44WebGuest58Xogium: good battery i guess, my main use is just listening to stuff though, i don't care for anything else really
17:34:28WebGuest58for example, on my ipod i use the failsafe theme (i think that's the one), minimal fluff just like i like it
17:35:04Xogiumright. Sounds like your ipod is perfect for this task then ? Unless I'm missing something here, that is ;)
17:35:20WebGuest58lol, you might be right
17:35:58Xogiumsince bluetooth isn't a possibility anyway at all in rockbox, there'd be no point in getting a bluetooth-capable device
17:42:20gnafuWebGuest58: I just got a Surfans F20 (Eros Q variant), and I'm happy with it so far. It has Bluetooth in the original firmware, and Rockbox is great for everything else.
17:42:45gnafuBut that's been my only foray into portable audio players since the Sansa Fuze many years ago ^^;.
17:43:09WebGuest58nice thanks
17:43:57gnafuI haven't actually tried out the stock firmware's Bluetooth functionality yet. I should do that soon, but I don't expect to use it much.
17:45:06WebGuest58looks very good and the price is right up my alley
17:59:23 Quit lebellium (Quit: Leaving)
18:00
18:02:01GaloisI also got the Hifi Walker H2 (Eros Q). In addition to bluetooth, the OF also supports USB-C audio, which I use. But the battery life is not super. It gets maybe 10-12 hours.
18:06:57WebGuest58awesome
18:15:10WebGuest58well guys i gotta go, thanks for now
18:15:21 Part WebGuest58
18:31:31***Saving seen data "./dancer.seen"
19:00
19:14:25 Quit jacobk (Quit: No Ping reply in 180 seconds.)
19:16:50 Join Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
19:28:20 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
19:37:39 Quit jacobk (Ping timeout: 260 seconds)
19:44:25 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
19:49:54 Quit jacobk (Ping timeout: 260 seconds)
19:55:02 Join massiveH [0] (~massiveH@2600:4040:a982:5400:bde8:5787:d30f:94cc)
19:55:16 Join _bilgus_ [0] (~bilgus@syn-162-154-213-134.res.spectrum.com)
19:56:49 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
19:59:12 Quit _bilgus (Ping timeout: 268 seconds)
20:00
20:00:14 Quit jacobk (Client Quit)
20:02:25 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
20:07:08_bilgus_the bin size jump 0o thats unexpected wth
20:07:42 Quit jacobk (Ping timeout: 276 seconds)
20:09:48 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
20:15:34 Quit jacobk (Ping timeout: 260 seconds)
20:20:20_bilgus_oh so it was actually pulling in the fallback one ugh
20:21:42 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
20:23:18_bilgus_pretty sure I just failed as I was figuring out a plce to put it but now I need to do those runtime tests again make sure I wasn't testing the fallback one but I can't imagine so
20:26:20_bilgus_Ah and the lower size was due to the removal of 50 ish format strings
20:27:51 Quit jacobk (Ping timeout: 276 seconds)
20:31:32***Saving seen data "./dancer.seen"
20:44:47 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
20:51:05 Join jacobk_ [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
20:51:44 Quit jacobk (Ping timeout: 260 seconds)
20:56:27 Quit jacobk_ (Ping timeout: 276 seconds)
21:00
21:08:00 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
21:11:35 Quit jacobk (Client Quit)
21:13:13 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
21:28:00 Quit jacobk (Quit: No Ping reply in 180 seconds.)
21:30:50 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
21:36:21 Quit jacobk (Quit: No Ping reply in 210 seconds.)
21:36:40 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
21:41:57 Quit jacobk (Ping timeout: 276 seconds)
21:50:28 Join jacobk [0] (~quassel@47-186-65-73.dlls.tx.frontiernet.net)
22:00
22:31:33***Saving seen data "./dancer.seen"
23:00
23:00:42 Join othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
23:13:43 Quit massiveH (Quit: Leaving)

Previous day | Next day