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 2020-12-11

00:30:32 Join saratoga [0] (
00:36:18 Quit ecs (Ping timeout: 260 seconds)
00:40:11saratogado we want to set a maximum album art size for performance reasons?
00:40:13 Join ecs [0] (
00:40:25saratogadecoding seems mostly dependent on file size not resolution
00:41:08saratogaa 12000x6000 1.7MB image took 9 seconds to load and resize on the Clip Zip
00:41:32saratogathe same image at 1/4th the pixels but 4x the file size took 18 seconds
00:42:12saratogafor performance reasons I'm tempted to say that files over 1MB or so should be rejected
00:42:34saratogaalthough that probably depends a lot on the device performance
00:43:11saratoganormally i wouldn't care, but playback is paused while the images decode, so if you have an enormous image on a slow player it can look like its frozen
00:44:02 Join alexbobp [0] (
00:44:24 Nick alexbobp is now known as Guest69432 (
00:44:57saratogag#3074 is seems to work on both device and sim, haven't managed to get a crash on either using lots of large album art
00:44:59fs-bluebot_Gerrit review #3074 at : Fix deadlocks when trying to buffer large album art. by Michael Giacomelli
00:48:08 Quit saratoga (Remote host closed the connection)
01:02:39braewoodsanother idea, make the loading cancelable if it takes too much time to finish
01:48:41***Saving seen data "./dancer.seen"
01:53:00_bilgus__saratoga I'd probably do Nx the player resolution or something
01:53:38_bilgus__like 128896 max is 1280x960
01:55:17_bilgus__speaking of do the results of the resize get saved or is this an every play cost?
02:19:32 Quit MrZeus (Ping timeout: 272 seconds)
03:10:26 Join prof_wolfff [0] (
03:17:37 Quit prof_wolfff (Ping timeout: 258 seconds)
03:48:43***Saving seen data "./dancer.seen"
04:18:17 Quit olspookishmagus (Quit: All for nothing)
04:51:25 Quit Stanley00 (Quit: Nice weekend)
05:01:30 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
05:07:07 Join blokdian [0] (abe0b209@
05:08:37blokdianso I was using the internet casually and then I was checking out a new FreshOS theme and I saw that I was banned
05:08:44blokdianI am using a guest account
05:09:05blokdianany help would be appreciated
05:11:34blokdianI am not banned from the website but just the forum, thought that would add some info
05:34:40 Quit blokdian (Ping timeout: 245 seconds)
05:36:08 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
05:48:45***Saving seen data "./dancer.seen"
06:03:40 Join blokdian [0] (abe0b209@
06:06:11 Quit blokdian (Remote host closed the connection)
06:11:10 Quit livvy (Remote host closed the connection)
06:23:25 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
06:41:42speachyblokdian: bans tend to only heppen when there are spammy links in the forum post.
06:41:58speachyunless you got caught up in an unrelated IP address ban.
07:00:06speachywhat was the username you used?
07:20:42braewoodsspeachy: erm. they're already gone.
07:21:14 Quit t0mato (Quit: Ping timeout (120 seconds))
07:21:19speachyyeah. but I did it for the lols. logs.
07:21:27 Join t0mato [0] (t0mato@gateway/vpn/mullvad/t0mato)
07:38:22 Quit akaWolf (Ping timeout: 260 seconds)
07:48:47***Saving seen data "./dancer.seen"
07:53:19 Join olspookishmagus [0] (
07:57:57 Join massiveH [0] (
07:58:01 Join MrZeus [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
08:48:21 Quit speachy (Quit: WeeChat 2.9)
09:05:13 Quit massiveH (Quit: Leaving)
09:48:51***Saving seen data "./dancer.seen"
09:49:14 Quit Strife89 (Quit: No Ping reply in 180 seconds.)
09:50:39 Join Strife89 [0] (
10:19:25 Quit beencubed (Quit: Leaving)
10:21:32 Join beencubed [0] (~beencubed@
10:21:57 Join saratoga [0] (
10:22:55saratoga_bilgus__: images are resized on every load, but normally it is so fast it doesn't really matter
10:24:53saratogaresolution is tricky to use since the ipods are the slowest players and they have reasonably large screens
10:25:35_bilgus__sure but that makes it that more important to limit to something reasonable
10:26:47saratogamaybe just an arbitrary 2 or 4MB limit
10:26:59saratogathats small enough that the user will realize that the big file loads, but it takes a while
10:27:11saratogathen hopefully reconsider using such big files
10:27:25_bilgus__maybe it can display a 'error' album cover
10:28:27_bilgus__big red X
10:41:18saratogathat might be a little annoying though, since the wps gets reformatted to show less info for a problem the user might not care about
11:11:08 Join J_Darnley [0] (
11:11:25 Quit jdarnley (Ping timeout: 258 seconds)
11:15:10 Join speachy [0] (~speachy@
11:16:01 Quit Rower (Ping timeout: 246 seconds)
11:21:48 Nick mendelmunkis is now known as mendel_munkis (
11:31:43 Quit livvy (Ping timeout: 240 seconds)
11:48:55***Saving seen data "./dancer.seen"
12:07:58 Join captainkewl [0] (
12:31:42 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
12:39:04 Quit mendel_munkis (Ping timeout: 240 seconds)
12:44:26 Join mendelmunkis [0] (
12:51:04 Join lebellium [0] (
13:11:27 Join akaWolf [0] (
13:48:59***Saving seen data "./dancer.seen"
14:00:51 Join Rower [0] (
14:48:24 Quit captainkewl (Remote host closed the connection)
15:00:30 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
15:14:41speachyblbro[m]: hey, I have another request for the voice generation in rbutil.
15:15:48speachyblbro[m]: it looks like voice-corrections.txt is compiled into the binary; could we add a URL to (optionally?) retrieve a more updated version?
15:17:45bluebrother^speachy: yes. I was also thinking about adding support for an updated rbutil.ini, i.e. use the internal one, and also try to download an updated version. Could use an url in build-info that points directly to cgit.
15:18:23bluebrother^though right now I was considering to finish the few minor issues left and then push 1.5.0.
15:18:32speachy(hmm, or maybe include in the zip file..)
15:18:54speachy(... though it looks like it's not used by anythin gother than rbutil...)
15:18:58bluebrother^then address things like that afterwards. As well as the ipodpatcher / sansapatcher elevation thing
15:19:56speachy_bilgus__: and it looks like the "invalid talk" thing isn't being generated properly in the zip files.
15:19:57bluebrother^which is something I'd like to get done in the not-too-distant-future.
15:20:43bluebrother^but given the updated bootloaders, player support and other internal changes a 1.5.0 before sounds like a good idea to me.
15:22:42bluebrother^so we don't delay it too much longer. And the patcher thing might take a bit.
15:22:50bluebrother^since we need to package things a bit differently then.
15:22:54speachybluebrother^: should the contents of be under apps/lang/* ?
15:23:58bluebrother^Yes. See voicefile.cpp:68
15:24:06bluebrother^well, not the contents, the zip file :)
15:24:45speachythe contents are in that directory too
15:25:03speachyso it would have to be extracted from the top-level .rockbox directory for the .vstrings to end up in apps/lang
15:27:52bluebrother^looking at the code it ignores the structure of the zip, and only tries to match files <language>.*
15:28:54bluebrother^ah. There's an additional talk file packaged,
15:29:33speachythat isn't supposed to be in there. I'm fixing that now
15:29:41bluebrother^but that's in the zip file, not in the folder
15:29:45bluebrother^ah, that explains things :)
15:36:51fs-bluebot_Build Server message: New build round started. Revision e4b514917e, 293 builds, 9 clients.
15:37:23speachyI'm not terribly happy with this fix but it's good enough until we end up with multiple InvalidVoice files.
15:37:57speachybut the fact that voice-corrections.txt isn't actually used by anything other than rbutil concerns me.
15:38:51speachyoh wait, there it is, in got lost in the noise
15:49:02***Saving seen data "./dancer.seen"
15:49:56fs-bluebot_Build Server message: Build round completed after 785 seconds.
15:49:59fs-bluebot_Build Server message: Revision e4b514917e result: All green
15:50:04fs-bluebot_Build Server message: New build round started. Revision 60139cf9f1, 293 builds, 9 clients.
16:01:25speachybluebrother^: so as far as voice-corrections.txt is concerned, I can put it into the file easily enough; I think that makes more sense than hardcoding it into the rbutil binary
16:02:18fs-bluebot_Build Server message: Build round completed after 734 seconds.
16:02:20fs-bluebot_Build Server message: Revision 60139cf9f1 result: All green
16:03:15bluebrother^hmm. Good point. Though I'm wondering if it would make more sense to download it, so we can have a more up-to-date version when creating the voice file. Though we could do both.
16:03:35bluebrother^but then otoh we could use that for creating talk files too. Not sure if we already do that.
16:03:52bluebrother^nope, we don't.
16:04:25bluebrother^but since this is for correcting words to the tts gets them pronounced properly we might want something like that for talk files too
16:04:52speachyif it's fetched, I don't think there's any point in not getting it directly via cgit or whatever
16:05:46speachyYou did say that you wanted voice generation to work without an internet connection
16:06:02speachyso that lends towards putting it into
16:07:28bluebrother^ah, right.
16:07:56bluebrother^well, we could have it in, and when we do have a working internet connection we can download an updated version via cgit.
16:08:07bluebrother^and do the same thing for rbutil.ini
16:08:35bluebrother^so we could add 2 more entries to build-info and simply have them urls point to cgit.
16:13:22speachyok, I'll add those. under [general] or [rbutil] or something else?
16:13:52speachygeneral.rbutilini_url and general.voicecorrections_url ?
16:14:26speachyor other.* instead?
16:18:17speachybluebrother^: It's in now, let me know if you want it changed.
16:18:58speachy(will get pulled into top-level build-info on the next build)
16:19:24fs-bluebot_Build Server message: New build round started. Revision 227ac2b715, 293 builds, 9 clients.
16:19:25bluebrother^looks good :)
16:19:40speachy(and this build has voice-corrections plopped into
16:19:58bluebrother^I'll work on updating first though. Really want to get things into a state where we can get rbutil 1.5.0 done
16:21:00speachyOnce that's out we should seriously consder a hwcodec-less 4.0 release
16:22:00speachybluebrother^: hmm, sapi_voice.vbs is also compiled in.
16:22:16bluebrother^yes, but that basically never changes.
16:22:38bluebrother^and its kinda part of the SAPI implementation.
16:22:39speachynot since 2012, anyway. :)
16:23:18bluebrother^it's simpler to call a vbs to call the SAPI stuff instead of reimplementing things ourselves. iirc that would need to use COM ...
16:23:58speachyquite true!
16:24:23fs-bluebot_Build Server message: Build round completed after 299 seconds.
16:24:26fs-bluebot_Build Server message: Revision 227ac2b715 result: 387 errors 0 warnings
16:24:43*speachy raises an eyebrow.
16:25:28speachyshould have done a clean build, d'oh.
16:27:05fs-bluebot_Build Server message: New build round started. Revision b339400531, 293 builds, 9 clients.
16:38:45fs-bluebot_Build Server message: Build round completed after 700 seconds.
16:38:49fs-bluebot_Build Server message: Revision b339400531 result: All green
16:53:22 Quit lebellium (Quit: Leaving)
17:36:56 Quit pamaury (Ping timeout: 240 seconds)
17:49:06***Saving seen data "./dancer.seen"
18:09:27_bilgus__sorry caps lock :P
18:10:17_bilgus__re invalid voice sorry gues sthat was a while ago
18:48:27 Join ac_laptop [0] (~ac_laptop@
19:28:50speachy_bilgus__: yeah, it's clear neither of us tested it on a pristine player. :)
19:37:32 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
19:49:08***No seen item changed, no save performed.
20:07:44_bilgus__see I
20:08:25_bilgus__'m pretty sure I tried that (what your patch has) originally)
20:11:11_bilgus__but maybe that was just because I screwed something up like that path
20:13:10speachyit was ending up in, with the full /path/to/build/directory/apps/lang in the zip file
20:13:14 Quit ufdm (Remote host closed the connection)
20:13:48 Join ufdm [0] (
20:16:50_bilgus__well good thinmg you were on the case sorry you had to fix my bug.
20:17:06speachyran into it while trying to check something else
20:17:14speachyone of those d'oh moments
20:33:23_bilgus__just saw 3037 I think from this point forward we should just get rid of all this depreciated stuff and the versioning all together
20:35:30_bilgus__maybe not the versioning persay but the strict ordering is no longer needed it should just be alphabetical at all times
20:35:36speachyI'd be okay with that, but I have been able to re-use un-deprecated strings in translations.
20:36:03speachythe language scripts sort everything alphabetically now
20:36:23_bilgus__yeah the whole rock translation thing is the major breaking point
20:37:07speachy(or do they just mirror the english language ordering? I've already forgotten, despite rewriting the damn thing twice..)
20:37:08 Join massiveH [0] (
20:39:00_bilgus__I'm just saying we have (the now working) invalid voice thing and build voices with every build the rules kinda are no longer so rigid
20:39:34_bilgus__ease of reading should be the only other concern now :p
20:40:19speachyok, it uses the english ordering. (well, it can techinally use any language; english is un-hardcoded now)
20:42:03speachythat reminds me, I still need to make the voice generation scripts parallelizable
20:42:09_bilgus__like adding new string at the end but I don't think some chunks for general versus plugins or other category (ies) is a bad idea either
20:42:51speachyI do like being able to logically group stuff together
20:43:00_bilgus__that and the user: field I'm not sure that was ever even used anywhere
20:43:06speachyyeah, it's not.
20:43:53_bilgus__do any of those voice engines use hints?
20:44:25_bilgus__like phonetic hints
20:44:51speachywell, there's the voice-corrections.txt stuff which I suppose fills that role
20:45:02_bilgus__yeah after a failure
20:45:40speachyfailure? it's just a pre-filter before passing the text to the TTS engine
20:45:57speachyor you mean falure requiring human correction
20:45:58_bilgus__ahh so you tweak them till they sound right?
20:46:09speachyin theory
20:46:15_bilgus__yeah the human part..
20:46:39speachyexcept for the japanese stuff I cut'n'pasted in just now, I think it's only actually used to expand some common acronyms
20:46:51_bilgus__that would be the only other field I might want to see in the list ratheer than user:
20:47:26speachythere's a bitrotten patch I saw that's intended to make CamelCase read out better
20:50:28speachyhuh,g#3079 is interesting
20:50:31fs-bluebot_Gerrit review #3079 at : list: reset viewport to avoid corrupting the text in the first line by Georg Gadinger
20:51:33_bilgus__that sounds like a bug on my part
20:54:03_bilgus__hmm no its set up right
20:54:21_bilgus__I'm gonna have to dig into this one
21:00:46_bilgus__yeah I'm missing how this is fixing the issue unless something like the scroll engine jumps in mid stroke
21:07:27speachyI saw somethign wonky today on my X3 −− I plugged in the player, it turned on, pulled up the USB action prompt
21:07:55speachyso far so good. except the prompt went away a moment later, and the prompt started scrolling
21:08:11speachybecause the underlying line was one that needed to scroll
21:08:44speachythe scroll thread stuff definitely has some wonkiness going on
21:15:46 Quit MrZeus (Ping timeout: 256 seconds)
21:33:07_bilgus__thats gotta be it
21:33:42_bilgus__though I'm not sure why we would be yielding to it without a sleep...
21:47:30_bilgus__struct viewport vp = *list_text_vp;
21:49:03_bilgus__I wonder what the actual copy does with that assignment
21:49:10***Saving seen data "./dancer.seen"
22:01:38 Quit cockroach (Quit: leaving)
22:06:33 Quit ac_laptop (Quit: WeeChat 3.0)
22:06:52 Quit ecs (Remote host closed the connection)
22:06:53 Join ac_laptop [0] (~ac_laptop@
22:06:59 Join ecs [0] (
22:30:45_bilgus__hmm still no closer to figuring that out the assignment works properly I figured maybe it was a pointer issue but no the only other thing I see that flipping the viewport would be marking the previous vp as dirty
22:31:28_bilgus__thus forcing a redraw but I'd expect that to happen as a matter of course anyway..
22:32:02_bilgus__guess I need to set up a situation where I can reproduce\
22:39:04 Quit ac_laptop (Ping timeout: 256 seconds)
22:47:01 Quit Topy44 (Ping timeout: 264 seconds)
22:53:08 Join Topy44 [0] (
23:30:37 Quit ungali (Quit: ungali)
23:31:00 Join ungali [0] (
23:31:00 Quit ungali (Changing host)
23:31:00 Join ungali [0] (~ungali@unaffiliated/ungali)
23:40:59 Nick Guest69432 is now known as alexbobp (
23:47:25 Quit TheSeven (Disconnected by services)
23:47:34 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
23:49:14***Saving seen data "./dancer.seen"

Previous day | Next day