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 2021-03-09

00:00:49_bilgusI'm not sure it matters anymore since we GEN lang files every build but it should ensure success
00:11:12chris_sI can do that. Wouldn't that be worse though? If I create a new id, all other languages will now see the english version instead whereas the current translations already seem reasonably close to what I'd change the original to if I just adjusted source and dest in the english (+ us language) file?
00:11:46_bilgusI think thats just going to make it more confusing
00:12:29_bilgusset the new ID and next translator of that lang will have the context to translate it properly
00:13:05chris_sok – should I mark the original id deprecated in the english.lang file only, or every lang file?
00:13:32_bilgusIt'd be in all of them
00:14:15_bilgusthen just english.lang for the new one (at the end)\
00:14:25chris_ssounds good
00:15:20_bilgusit was the old way we did it but like I said with lang GEN running each round maybe versioning isn't so very important now
00:37:51braewoods_bilgus: go figure. battery strength doesn't work under windows. no idea why.
00:38:22braewoodsguess i'll look into plan B
00:38:45_bilgusdo other things do battery level without a custom driver?
00:39:06braewoods_bilgus: i suspect they do, bluetooth devices supposedly report battery level under recent win10
00:39:06_bilgus(in windows)
00:39:19braewoodsbut win10 is closed source so i have no idea how to debug an issue like this
00:39:56_bilguswell if its old enough to be in NT theres win2000 source floating about
00:40:04braewoodsi was testing on win10
00:40:18braewoodsand afaik battery strength wasn't used much until 2010+
00:40:24braewoodsat least on Linux
00:40:32braewoodsi think i know what will work though
00:40:36braewoodsHID Power
00:40:46braewoodsthe problem is it requires its own endpoint and interface
00:40:53_bilgushard telling if MS would have even had the need way back when
00:41:15braewoodsi'll try a test driver
00:41:19braewoodsbut later
00:41:33braewoodsit's more work than battery strength but if the results aren't cross platform
00:41:36_bilgusI only mention it because i've gone spelunking through that source looking at weird quirks
00:42:13braewoodsi want to try to make this work on the common hosts at least
00:42:24braewoodsbut windows doesn't seem too standard
00:42:26braewoodsD
00:42:28braewoodsD:
00:42:52_bilgusthey are their own standard?
00:42:56_bilguslol
00:43:00braewoodsbasically
00:43:14braewoodsi was hoping they would use this simple method
00:43:17braewoodsor notice it
00:43:18braewoodsnope
00:43:31braewoodswelp, battery strength was more of a hack in our case anyway
00:43:41braewoodsHID Power is more appropriate
00:43:48braewoodsprovided it works
00:43:51braewoodsugh
00:43:51_bilgusonly thing I noted worse on linux is PLUSE Audio vs WASAPI
00:43:58braewoodspulse
00:44:17_bilgusbut there are promises to change that soon
00:45:09braewoodsfor that matter we could keep both BS and HP
00:45:19braewoodsdo dual reporting
00:45:27CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
00:45:27*braewoods shrugs
00:45:30braewoodsanyway
00:45:37braewoodsi'll see if this works better
00:45:47braewoodsBS was good practice in any case
00:45:51braewoodseven if windows doesn't seem to like it
00:47:34_bilguswhy not just a standalong logger for win?
00:47:39_bilgusalone
00:48:25_bilgusthen you can just do it the linux way and let windows come along for the ride
00:48:28braewoodsstandalone logger?
00:49:00_bilgusintermediary program
00:49:11braewoodsi wouldn't know how to do that on windows
00:49:21braewoodsassuming it can even find the BS output
00:50:02_bilgusit should be able to find endpoints but perhaps a driver is still in order
00:50:24braewoodsno idea
00:50:31braewoodsanyway, i assumed windows would monopolize them
00:50:45_bilgusand in that case meh driver signing is even more strict in W10 and they want too damn much
00:51:04braewoodsyes, which is why i prefer to use a generic driver
00:51:23braewoodsbut it's hard to tell what generic features windows 10 supports
00:51:37braewoodsso i'll just try a basic hid power driver and see what windows does
00:51:53_bilguswonder if their linux subsystem is mature enough to just work
00:52:12braewoodsi doubt it would work for this
00:52:21braewoodseven so, i don't want the end user to have to install anything for this to work
00:52:24braewoodsif i can help it
00:53:18braewoodsi'll slap together a test driver and see what happens
00:53:33braewoodsi need the practice with RB's usb stack anyway
01:00
01:06:37 Quit p0x (Quit: Connection closed)
01:07:27_bilgusbraewoods, windows 2000 had features for portable device battery status
01:07:57_bilgusit was or is in with the systray exe originally
01:11:24_bilgusah tht sounds a lot like the one for UPS or maybe the laptop one maybe battery powered portables is LAptops in 1998 speak
01:19:27***Saving seen data "./dancer.seen"
01:23:31 Quit chris_s (Quit: Connection closed)
01:24:33 Join chris_s [0] (5fdf48e9@ip-95-223-72-233.hsi16.unitymediagroup.de)
01:26:47chris_sbilgus: g#3199
01:26:49fs-bluebotGerrit review #3199 at https://gerrit.rockbox.org/r/c/rockbox/+/3199 : UI: Rename "Play Next" to "Clear List & Play Next" by Christian Soffke
01:29:29 Quit chris_s (Quit: Connection closed)
01:42:59_bilguslooks good
02:00
02:00:59 Quit koniu (Remote host closed the connection)
02:01:24 Join koniu [0] (~koniu@gateway/tor-sasl/koniu)
02:07:18 Join lebellium [0] (~lebellium@89-92-69-66.hfc.dyn.abo.bbox.fr)
03:00
03:19:31***Saving seen data "./dancer.seen"
04:00
04:11:03 Quit S|h|a|w|n (Read error: Connection reset by peer)
04:13:04 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
04:23:29 Join Rower- [0] (~Rower@138.199.59.73)
04:24:38 Quit Rower (Ping timeout: 264 seconds)
05:00
05:08:24 Quit Lonoxmont (Ping timeout: 245 seconds)
05:13:26 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
05:17:14 Quit mendelmunkis (Remote host closed the connection)
05:19:32***Saving seen data "./dancer.seen"
05:20:19 Quit ArsenArsen (Ping timeout: 265 seconds)
06:00
06:02:46speachychris_s, _bilgus, with the "recent" genlang stuff there's no point in mucking with "deprecated" language strings any more. Either they're used, or they're not.
06:04:18speachyit's bad form to semantically change the "meaning" of LANG_xxxx but the translation site does flag that scenario as a problem that needs addressing.
06:05:23speachyso it's actually better that you don't mark the old string as deprecated in english; just nuke it altogether.
06:05:55speachyand leave the other languages alone; the unused string will get dropped automagically.
06:12:44speachywe've been getting a lot stricted with respect to voice files matching up with the main binaries.
06:12:51speachy*stricter
06:16:47speachymost of the lang files have a pile of no-longer-present strings in them. It really comes down to how recently they were updated.
06:18:29fs-bluebotBuild Server message: New build round started. Revision 3b9a803a5b, 293 builds, 9 clients.
06:30:22fs-bluebotBuild Server message: Build round completed after 714 seconds.
06:30:25fs-bluebotBuild Server message: Revision 3b9a803a5b result: All green
06:30:28fs-bluebotBuild Server message: New build round started. Revision 5c3bbc6a9c, 293 builds, 9 clients.
06:31:15speachynow what happened here is a loophole in our checking −− we remove one string and add one, leaving the overall enumeration the same, so leaving the old one as 'deprecated' would be a good idea
06:35:16 Quit Guest48842 (Quit: All for nothing)
06:36:00 Join olspookishmagus [0] (~pookie@snf-137798.vm.okeanos.grnet.gr)
06:41:18fs-bluebotBuild Server message: Build round completed after 650 seconds.
06:41:20fs-bluebotBuild Server message: Revision 5c3bbc6a9c result: All green
06:41:21fs-bluebotBuild Server message: New build round started. Revision 6387d0352e, 293 builds, 9 clients.
06:45:38braewoods_bilgus: yea..
06:45:48braewoods_bilgus: we just want periphereal battery reporting
06:49:39_bilgusSpeachy, So still important then?
06:49:59speachyonly in the sense that two builds with the same number of strings will think the voice file is kosher
06:50:59speachyI'd like to implement some sort of checksum (seems like we went around this topic a few times already) to make it stricter
06:52:01speachyso checksum the LANG_xxx identifiers themselves, and include that in the build and the voice header?
06:52:44fs-bluebotBuild Server message: Build round completed after 684 seconds.
06:52:45speachy(or perhaps checksumming the english source strings is the better option, to guard against possible semantic changes)
06:52:46fs-bluebotBuild Server message: Revision 6387d0352e result: All green
06:52:47fs-bluebotBuild Server message: New build round started. Revision d0ab6b24dc, 293 builds, 9 clients.
06:54:35speachybut regardless of 'deprecated' being arguably necessary for some scenarios, there's no point in touching the non-english language files to add deprecated tags or remove now-unused strings.
06:55:18 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
06:57:40speachythe translate.rockbox.org site will clean that up when the translators update things.
07:00
07:00:30speachythe tooling will automatically nuke any string in the non-english languages if it's empty (or 'none') in English.
07:01:17speachy(which, come to think of it, might not be what we want in all situations; it's possible for some generic construct to be needed for certian non-english languages, but I have no idea if we implement anyting like that...)
07:04:25fs-bluebotBuild Server message: Build round completed after 698 seconds.
07:04:27fs-bluebotBuild Server message: Revision d0ab6b24dc result: All green
07:10:14speachy(oh, by "tooling" I mean the updatelang/genlang stuff used in the main build and voicefile generation)
07:10:53speachy(t.r.o implements the same logic for flagging things to present to translators)
07:19:34***Saving seen data "./dancer.seen"
07:37:06 Join ufdm [0] (~ufdm@c-73-164-63-214.hsd1.mn.comcast.net)
07:56:40 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
08:00
08:06:15 Nick vup2 is now known as vup (~~~~@46.101.193.235)
08:25:43_bilgusOK cool, yeah checksum over the voice strings and store it in the header that should be pretty robust
08:26:58fs-bluebotBuild Server message: New build round started. Revision f348d921b6, 293 builds, 8 clients.
08:39:23braewoodsinteresting
08:39:25fs-bluebotBuild Server message: Build round completed after 748 seconds.
08:39:28fs-bluebotBuild Server message: Revision f348d921b6 result: All green
08:39:31braewoodsseems battery strength can be delivered 2 ways
08:39:41braewoodseither solicited or unsolicited
08:39:50braewoodswe do unsolicited... maybe windows doesn't like that?
08:43:27braewoodsspeachy: i'll try an experiment.
08:43:53braewoodsi'll see if i can switch the battery reporting to on demand
08:44:12braewoodsno idea if it'll work
08:44:13braewoodsbut
08:44:19braewoodsthat may be the issue
08:48:53_bilgusI really don't see mention of battery reporting on windows in general lots about how w10 can do bluetooth battery levels and UPS stuff but very scant on 'other devices' battery levels
08:49:15braewoodsindeed
08:49:35braewoodsbut there's evidence of digitizers working
08:49:38braewoodsso
08:49:44 Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
08:49:48braewoodsno idea as of yet
08:50:13braewoodsimplementations can be quite picky
08:50:30braewoodsi suspect Linux happens to work with both for the sake of compatibiltiy
08:50:38_bilgusmight wanna ask on the new old thing blog −− Raymond Chen
08:50:41braewoodsbut maybe windows only works with one method
08:50:58braewoodseither way seems there's 2 ways to report battery strength
08:51:06braewoodsdepending on how it is sent
08:52:07speachyok, redoing the voice file to include a checksum will be an issue; it's going to require coordination with rbutil
08:52:59speachyand that has no awareness of the origin LANG_xx identifier or any checksums.
08:53:10speachyor the source strings I meant
08:56:12speachyHmm. Could just checksum the engligh.vstrings file instead, and embed that CRC into the binary and voicefiles.
08:58:50 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
09:00
09:02:51_bilgusmaybe set it up as a dummy voice_id like Lang_Voice_ver and have it spit out the ver.crc or something
09:02:55speachyyeah
09:19:02 Quit jdarnley (Ping timeout: 256 seconds)
09:19:37***Saving seen data "./dancer.seen"
09:20:05 Quit ender| (Quit: CLOUD: Computation Located On Unknown Device)
09:25:59 Join J_Darnley [0] (~J_Darnley@d51A44418.access.telenet.be)
09:38:02 Join ender| [0] (~ender1@2a01:260:4094:1:6045:6fff:fedd:cbf3)
09:49:42fs-bluebotBuild Server message: New build round started. Revision f4b891e408, 293 builds, 9 clients.
09:58:20 Nick edhelas is now known as edhelas_ (9d94237298@2a03:4000:51:f44:4e1:2ff:fe00:4257)
10:00
10:03:01fs-bluebotBuild Server message: Build round completed after 799 seconds.
10:03:04fs-bluebotBuild Server message: Revision f4b891e408 result: All green
10:10:59fs-bluebotBuild Server message: New build round started. Revision b2d13f1793, 293 builds, 9 clients.
10:14:56braewoodsspeachy: how can i debug usb?
10:15:30braewoodsi think i found another option for checking battery strength
10:15:37braewoodsbut need to debug stuff
10:16:15speachybraewoods: running this as a vm or native under windows??
10:16:34speachyunder linux you can use the built-in usbmon stuff (using wireshark or tcpdump or whatever)
10:16:46speachyit also works to capture usb stuff passed throughto a VM
10:17:07speachyudner windows natively (or even as a VM) usbpcap is awesome. it's bundled with the windows wireshark builds
10:17:30speachythat'll give you the packet-level view of what's going on. if you need electrical/bus-level views that's going to take specialized hardware.
10:17:47speachy(And for what you're doing, packet-level debug is plenty sufficient)
10:19:22 Quit massiveH (Quit: Leaving)
10:29:35fs-bluebotBuild Server message: Build round completed after 1115 seconds.
10:29:37fs-bluebotBuild Server message: Revision b2d13f1793 result: All green
10:52:51 Quit Saijin_Naib (Read error: Connection reset by peer)
11:00
11:09:41 Quit tomato (Quit: Ping timeout (120 seconds))
11:18:47 Join Saijin_Naib [0] (~Saijin_Na@2603-7081-1d05-7230-3cb9-9e28-e430-269e.res6.spectrum.com)
11:19:39***Saving seen data "./dancer.seen"
11:20:31 Join tomato [0] (t0mato@gateway/vpn/mullvad/tomato)
12:00
12:25:48fs-bluebotBuild Server message: New build round started. Revision 4e6609e9f2, 293 builds, 9 clients.
12:37:29 Join ArsenArsen [0] (~Arsen@fsf/member/ArsenArsen)
12:40:26fs-bluebotBuild Server message: Build round completed after 878 seconds.
12:40:28fs-bluebotBuild Server message: Revision 4e6609e9f2 result: All green
13:00
13:19:41***No seen item changed, no save performed.
13:27:27speachybraewoods: if you want other things to muck with, you could also create a USB UART/ACM endpoint to use as a debugging console. :)
13:35:23 Join jdarnley [0] (~J_Darnley@d51a44418.access.telenet.be)
13:36:45 Quit J_Darnley (Ping timeout: 246 seconds)
14:00
14:23:23 Quit daswf852 (Read error: Connection reset by peer)
14:26:31 Join daswf852 [0] (~daswf852@unaffiliated/dwf)
14:26:35 Quit daswf852 (Read error: Connection reset by peer)
14:34:28 Join daswf852 [0] (~daswf852@unaffiliated/dwf)
14:40:00 Quit koniu (Remote host closed the connection)
14:40:32 Join koniu [0] (~koniu@gateway/tor-sasl/koniu)
14:43:15 Quit Saijin_Naib (Disconnected by services)
14:43:19 Join Saijin-Naib [0] (~Saijin_Na@2603-7081-1d05-7230-3cb9-9e28-e430-269e.res6.spectrum.com)
14:43:23 Quit daswf852 (Quit: ZNC 1.8.2 - https://znc.in)
15:00
15:01:10 Join J_Darnley [0] (~J_Darnley@d51A44418.access.telenet.be)
15:03:33 Quit jdarnley (Ping timeout: 264 seconds)
15:10:50 Quit Saijin-Naib (Ping timeout: 264 seconds)
15:12:37 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
15:18:02 Join MrZeus [0] (~MrZeus@2a02:c7f:a0aa:4400:2d12:9195:f056:c03d)
15:19:42***Saving seen data "./dancer.seen"
15:20:10 Quit MrZeus (Read error: No route to host)
15:23:15 Join MrZeus [0] (~MrZeus@2a02:c7f:a0aa:4400:34e5:7d2a:4362:8430)
15:24:01 Join MrZeus_ [0] (~MrZeus@89.238.130.71)
15:28:03 Quit MrZeus (Ping timeout: 260 seconds)
15:29:24 Quit lebellium (Read error: Connection reset by peer)
15:29:44 Join lebellium [0] (~lebellium@89-92-69-66.hfc.dyn.abo.bbox.fr)
15:30:21 Join amachronic [0] (5284b8a3@82.132.184.163)
15:31:28amachronichi, I just updated my FiiO port, it's in much better shape now and it actually boots!
15:32:25amachronicalso pushed a fix for another MIPS-specific problem which should probably be merged
15:34:09fs-bluebotBuild Server message: New build round started. Revision 777f63d529, 293 builds, 9 clients.
15:34:24speachymips fix committed
15:36:49amachronicspeachy: thanks; when you've got the time I was wondering what you'd like to see done before merging the M3K port.
15:38:23speachyIt's probably worth merging as-is, but before we can start providing builds/etc there has to be a sane way to flash our bootloader (ideally with fallback to the OF) and restore factory loadout.
15:41:50amachronicthe update I pushed has an installer which can flash RB or restore the OF from the debug menu, but dual boot isn't working yet. the installer could probably use a bit of improvement too.
15:41:51speachythere's only a handful of missing features vs the OF; USB dac, recording, and 24-bit audio
15:42:27speachyhonestly the OF is such a janky mess that I can't see anyone wanting to actually _use_ it vs rockbox. :D
15:42:42amachronicdo you have an M3K yourself?
15:44:04speachyyeah, I have one here.
15:44:17speachy(oh, one other feature rockbox lacks −− ExFAT support)
15:44:58amachronicmultiple filesystem support is a huge can of worms, unless exfat is easily built on top of what we have... (shudders)
15:45:31speachyit's not _that_ bad, beyond the complexities of exfat itself
15:45:47speachywe do have mostly sane layering going on
15:46:19amachronichaven't looked hard enough I guess, then... I just remember seeing calls to FAT FS sprinkled around in places.
15:47:14amachronicas for 24-bit audio: isn't Rockbox limited to 16 bits internally?
15:48:12speachycurrently, yes. we do handle higher sample rates just fine though, assuming the raw oomph is there.
15:49:18fs-bluebotBuild Server message: Build round completed after 908 seconds.
15:49:20fs-bluebotBuild Server message: Revision 777f63d529 result: All green
15:49:38amachronicon another point: I was also wondering how you go about adding new options. The settings code looks like a bit of a mess to me and I don't know how to properly update the LANG strings etc.
15:50:00speachyyeah, it's complicated.
15:50:37amachronicseems you guys have been confused about it too the past few days?
15:52:20speachythere's the raw "setting" definition (type, name, enum/range, etc), there's the location in RAM where it goes (struct global_settings ?)
15:52:29speachyand where you put it into the menu
15:52:56speachyand what happens when it is set/changed (just update the variable, call a function, etc)
15:53:19speachythe most recent confusion was that some of those didn't share the sme set of #ifdefs wrapping things
15:53:47speachyan unrelated issue was a broken language file that uncovered a couple other bugs in the tooling
15:56:13amachronicI'll try adding the settings then, the DAC's filter roll off values need to be put in and there's a power saving function I wanted to add too.
15:57:19speachythere are already some rolloff options (LANG_FILTER_ROLL_OFF, _FILTER_SHARP, _FILTER_SHOW, FILTER_SHORT_SHARP, FILTER_SHORT_SHOW, FILTER_SHORT)
15:58:19speachylook for AUDIOHW_HAVE_FILTER_ROLL_OFF in settings.c
15:58:26amachronichmm, last time I checked I thought there was something that didn't match up... I thought one of the values was "bypass"
15:58:53amachronicam I right to assume new strings only go in the "english" lang file? or do I have to update every lang file as well?
15:58:54speachythere's more than one set
15:59:14speachyonly english. Unless you're inclined to translate it into any other languages you're familiar with
15:59:53speachythere are three different sets of rolloff filter options for different hw. maybe there will end up bieng a fourth?
16:00
16:01:27amachronicwill have to check the rolloff settings more carefully, and if there's a matching set I'll just re-use it.
16:02:13speachyif there isn't, it can be hacked in using the same "strategy"... or refactoring if feasible.
16:03:48amachronicno, it's definitely changed since I last looked and what I need is there, (AUDIOHW_HAVE_SHORT2_ROLLOFF). must one of those recent linux-based ports added it.
16:04:05speachyyeah, I think I added that for the X3ii/X20
16:08:32amachronicI wanted to add a function in audiohw.h for enabling power saving modes. The M3K's DAC can cut power use by 2/3rds in its "low mode" vs the high mode
16:08:58amachronic(or allegedly that's the savings, I didn't test if it's true)
16:09:41speachyyou got rid of HAVE_CPU_CACHE_ALIGN?
16:10:39amachronicI think I redefined it globally in system.h for all MIPS
16:12:06amachronicjust checked and that's on master already (thought so)
16:12:09speachyah, I see it. cool.
16:12:43amachronicspeaking of that, did you have any luck getting the FAT alignment code up and running?
16:13:30 Quit alexbobp (Ping timeout: 272 seconds)
16:13:31speachyI got it running and it appeared to work... until it didn't, though that particular garble-garble-crash could have been unrelated (I had a bunch of other flaky stuff built in)
16:15:15amachronicif gerrit is up to date with your code, then one problem is you're using the "rc" code from storage_read_sectors wrongly
16:15:27 Join Guest59597 [0] (~alex@meowface.org)
16:15:47amachronicI checked the existing ports before doing mine, and "rc" is just a typical "zero on success nonzero on failure" code
16:16:41speachywell, that explains a few things. :D
16:17:22 Quit pamaury (Ping timeout: 265 seconds)
16:18:40amachronicso if you change it to advance by "count" sectors each time I suspect it would otherwise work
16:24:23speachyrunning it now. let's see how it goes during my driving tomorrow. (need to regen the voice files though)
16:27:57braewoodsspeachy: how does one use logf?
16:28:09braewoodsi can't seem to get anything to log to the buffer or so
16:28:14amachronic#define LOGF_ENABLE
16:28:15braewoodsi enabled the logf option for build
16:28:30amachronicor rather #define LOGF_ENABLE before #include "logf.h"
16:28:35speachywas about to ask
16:28:46 Quit Guest59597 (Ping timeout: 265 seconds)
16:28:51braewoodsi wanted to debug the rockbox side of things
16:29:07braewoodsanyway i found out Linux checks for battery strength in two places
16:29:12braewoodsinput fields and feature fields
16:29:22 Join alexbobp_ [0] (~alex@meowface.org)
16:29:25speachyamachronic: define it at the very top of the file, as logf.h might get included via another #include
16:29:55braewoodsspeachy: so i have to manually enable logf where i need it?
16:29:58braewoodsit's not global?
16:31:40amachronicbraewoods: it is enabled only on a per-file basis, I guess so you don't drown in the logf output from other files
16:31:49braewoodsok...
16:32:06amachroniclike speachy said it is probably best to define LOGF_ENABLE at the top of the file
16:32:35speachybraewoods: it's a limiited-size ring buffer in RAM, very easy to drown it.
16:33:41braewoodsok
16:34:19braewoodsi'll need to use it, see what the host is doing
16:34:24speachyamachronic: have you given any further thought to the keymap situation?
16:34:29 Quit alexbobp_ (Remote host closed the connection)
16:35:56 Join alexbobp_ [0] (~alex@meowface.org)
16:36:14 Join Rower [0] (~Rower@78-73-72-39-no2340.tbcn.telia.com)
16:36:47 Quit Rower (Client Quit)
16:38:56speachyalso, it seems that putting the stack in IRAM would be a good idea for jz47xx too
16:38:57 Quit Rower- (Ping timeout: 264 seconds)
16:39:11amachronicspeachy: that's where I got the idea :)
16:39:23speachyaaah, outsmarted myself. :D
16:40:50braewoodsspeachy: kinda wish i could built a stripped down RB just to make build testing faster
16:40:57braewoodssince i don't need parts of it
16:41:31speachybraewoods: set plugins=no in tools/configure for your target, that'll speed up a lot of stuff.
16:41:48speachyotherwise, as the saying goes, there's no replacement for displacement...
16:42:01speachy(ccache helps, if you have it installeD)
16:44:03braewoodsah, i meant the installation mainly
16:44:07braewoodsrb isn't exactly fas
16:44:40braewoodsi switched BS to feature reporting
16:44:45braewoodstime to see what happens
16:44:56braewoodsi'll see if i get any requests from the host
16:46:05braewoodsyep we're getting them now
16:46:12braewoodsnow need to handle them
16:46:33speachyamachronic: I'm going to have to come back to this later tonight. My intent is to commit it unless I run into some sort of WTF.
16:47:39amachronicspeachy: no problem, there's no need to rush it anyway
16:48:54amachronicas for the keymaps, I should be able to mimick the OF's behavior enough to re-use the keymaps from the Linux port if you want. But currently I'm doing an i2c driver overhaul so I can't get around to it just yet.
16:50:41speachyalso for the record, when using the X3's OF, don't plug in headphones when you have the disk mounted.
16:51:42braewoodsspeachy: why? you can hear it moaning in agony?
16:51:46braewoods:p
16:52:05speachyit'll disconnect from the host, and go into the menus, ignoring the state of the keylock switch
16:52:30braewoodsi see
16:52:45 Quit alexbobp_ (Ping timeout: 264 seconds)
16:53:35 Join alexbobp_ [0] (~alex@meowface.org)
16:54:21speachy(looks like a couple different OF bugs stacked up. but that's what you're stuck with when you can't trust what rockbox is reading/writing to storage...)
16:58:28 Quit alexbobp_ (Ping timeout: 260 seconds)
16:59:12 Join alexbobp_ [0] (~alex@meowface.org)
16:59:18 Quit lebellium (Quit: Leaving)
16:59:21braewoodssounds like something MTP could help with
17:00
17:01:49speachywell, that requires even more trust in rockbox. :)
17:05:01braewoodsinteresting, linux requests a feature report of different size than the original report
17:05:02braewoodsgo figure
17:05:19braewoodsit's probably wanting a whole 32 bit integer
17:07:15braewoodstime to see what windows does
17:09:14 Quit alexbobp_ (Ping timeout: 260 seconds)
17:09:32 Join alexbobp_ [0] (~alex@meowface.org)
17:11:05 Join ac_laptop [0] (~ac_laptop@186.2.247.129)
17:15:43 Quit alexbobp_ (Remote host closed the connection)
17:15:52 Join alexbobp [0] (~alex@meowface.org)
17:16:14 Nick alexbobp is now known as Guest22103 (~alex@meowface.org)
17:19:43***Saving seen data "./dancer.seen"
17:20:10 Join Saijin_Naib [0] (~Saijin_Na@2603-7081-1d05-7230-280f-3789-ab28-3929.res6.spectrum.com)
17:21:50 Quit Guest22103 (Ping timeout: 260 seconds)
17:22:15braewoodsspeachy: huh... windows doesn't even acknowledge the feature report exists.
17:22:22braewoodsi get no requests
17:22:26*braewoods boggles
17:28:54braewoodswell, this explains a lot
17:29:03braewoodshttps://docs.microsoft.com/en-us/windows-hardware/drivers/hid/hid-architecture
17:29:13braewoodsusage page 0x06 is not mentioned
17:32:56braewoodsbaffling
17:37:38 Join alexbobp_ [0] (~alex@meowface.org)
17:42:08 Quit alexbobp_ (Ping timeout: 245 seconds)
17:44:06 Join alexbobp_ [0] (~alex@meowface.org)
17:49:45 Quit alexbobp_ (Ping timeout: 264 seconds)
17:50:37 Join alexbobp_ [0] (~alex@meowface.org)
17:53:45 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
17:56:43 Quit alexbobp_ (Ping timeout: 245 seconds)
17:59:50 Join alexbobp_ [0] (~alex@meowface.org)
18:00
18:14:29 Quit amachronic (Quit: Connection closed)
18:18:29 Quit alexbobp_ (Ping timeout: 265 seconds)
18:18:38 Join alexbobp [0] (~alex@meowface.org)
18:19:00 Nick alexbobp is now known as Guest90807 (~alex@meowface.org)
18:23:36 Quit Guest90807 (Ping timeout: 256 seconds)
18:39:53 Join alexbobp_ [0] (~alex@meowface.org)
18:55:20braewoodsspeachy: i think i'm just going to rework how we do battery reporting to use HID feature reports and call it a day
18:56:34braewoodsi discovered MTP provides another way to report battery levels too
18:56:43braewoodswe'll keep this for a basic method
19:00
19:03:07 Quit pamaury (Ping timeout: 272 seconds)
19:18:40 Quit shapeless (Ping timeout: 258 seconds)
19:19:46***Saving seen data "./dancer.seen"
19:41:38 Quit ac_laptop (Quit: WeeChat 3.0)
19:42:00 Join ac_laptop [0] (~ac_laptop@186.2.247.129)
19:42:23braewoodsugh
19:42:33braewoodscaused a usb stack crash
19:42:40braewoodsand i can't reboot theplayer
19:42:43braewoodsoh well time to wait it out
19:48:14braewoodsone of the perils when you're not familiar with the system you're using
20:00
20:34:55 Join ats_ [0] (~ats@cartman.offog.org)
20:38:26 Join MulX [0] (~mulx@eva.aplu.fr)
20:38:59 Quit atsampson (*.net *.split)
20:38:59 Quit M-iam-some0nee[m (*.net *.split)
20:39:00 Quit APLU (*.net *.split)
20:39:03 Nick MulX is now known as APLU (~mulx@eva.aplu.fr)
20:44:27 Quit MrZeus_ (Ping timeout: 246 seconds)
21:00
21:01:23 Join M-iam-some0nee[m [0] (m-iam-some@gateway/shell/matrix.org/x-wuuvpjzawbblesjh)
21:19:49***Saving seen data "./dancer.seen"
21:22:34 Quit kugel_ (Ping timeout: 272 seconds)
21:28:14 Join kugel [0] (~kugel@rockbox/developer/kugel)
21:39:33 Quit ac_laptop (Ping timeout: 264 seconds)
21:56:54braewoodsspeachy: i noticed linux always asks for 4 bytes of data even though my featur erport only has 1 byte of data to return so
21:57:03braewoodsi decided to extend it with padding bytes
21:57:10braewoodsuse the extra for reserved fields or so
21:57:16braewoodsi already found a use for 1
21:57:27speachyyou can ask for more, there's no guarantee that you'll get it
21:57:51braewoodstrue but given how it is written
21:58:04braewoodsi thought i'd put the RB model number in the first byte
21:58:16braewoodscould be useful for probing for the connect devices
21:58:23speachy(when you post the request URB you only tell it the maximum size you can accept; shorter reads are quite legal per the USB spec. Now the HID spec might beg to differ..)
21:59:05speachyif we're talking about RB models, IMO we're better off just reporitng that via the iManufacturer/iModel strings
21:59:27braewoodshttps://github.com/torvalds/linux/blob/master/drivers/hid/hid-input.c#L355
21:59:44braewoodsgiven how it is written
22:00
22:00:16braewoodsit then reads from the second byte
22:00:25braewoodswhich it treats as a percent from what i can tell
22:00:50braewoodsoh
22:00:54braewoodsit only requires 2
22:01:06braewoodsi see, fair enough
22:01:18braewoodsugh usb is confusing
22:07:22braewoodsspeachy: in any case, i think switching battery strength to this mode has advantages.
22:07:37braewoodsnamely, we can just let the host decide when it wants the data
22:08:13braewoodsperhaps we can also probe it ourselves for other OS with no support
22:14:49braewoodsspeachy: do we have a rockbox icon we could use on the host side?
22:15:00braewoodsseems MTP also has fields for custom icons
22:15:08braewoods"device icon"
22:15:13speachyyes, but what format?
22:15:17braewoods.ICO
22:16:11speachythere's always the favicon.iso off of the rockbox site
22:16:18speachyhttps://www.rockbox.org/favicon.ico
22:16:31braewoodsah
22:16:36braewoodsif we had another i could try to
22:16:40braewoodshm
22:16:44braewoodsoptimize it for embedding
22:16:51braewoodsthough i dunno how useful it'd be
22:17:02braewoodsmtp has a lot of fields not necessarily useful
22:27:51 Join f1reflyylmao [0] (~f1refly@2a01:c23:9430:c900:9d33:156d:5e4f:9e24)
22:28:38 Quit cockroach (Quit: leaving)
22:29:48 Quit f1refly (Ping timeout: 260 seconds)
22:29:49 Nick f1reflyylmao is now known as f1refly (~f1refly@2a01:c23:9430:c900:9d33:156d:5e4f:9e24)
23:00
23:19:53***Saving seen data "./dancer.seen"

Previous day | Next day