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 2021-06-05

01:00
01:25:54***No seen item changed, no save performed.
01:42:40 Quit jschwart (Ping timeout: 272 seconds)
01:49:22 Join jschwart [0] (~quassel@2001:985:2c6e:0:b00b:32ff:fe28:5567)
01:56:58 Join daswf8527 [0] (~daswf852@user/daswf852)
01:59:42 Quit daswf852 (Ping timeout: 264 seconds)
01:59:42 Nick daswf8527 is now known as daswf852 (~daswf852@user/daswf852)
02:00
02:17:40 Quit _vanadal ()
02:29:31 Join K4os_ [0] (~K@ip5f5af5c5.dynamic.kabel-deutschland.de)
02:37:32 Join lebellium [0] (~lebellium@lfbn-str-1-613-4.w90-126.abo.wanadoo.fr)
02:53:40 Quit K4os_ (Ping timeout: 244 seconds)
03:00
03:25:56***Saving seen data "./dancer.seen"
05:00
05:13:17 Join johnb2 [0] (~johnb2@p4fd10b03.dip0.t-ipconnect.de)
05:25:58***No seen item changed, no save performed.
05:47:52 Join ideasman42 [0] (~ideasman4@2402:b801:2844:2100::1)
05:48:37 Quit ideasman42 (Client Quit)
05:48:51 Join ideasman42 [0] (~ideasman4@2402:b801:2844:2100::1)
05:51:24 Nick TorC is now known as Padauk (~Tor@user/torc)
05:53:14 Nick Padauk is now known as Marqueteur (~Tor@user/torc)
05:53:25 Nick Marqueteur is now known as Amboyna (~Tor@user/torc)
05:53:31 Nick Amboyna is now known as TorC (~Tor@user/torc)
05:53:33 Quit ideasman42 (Client Quit)
05:53:53 Join ideasman42 [0] (~ideasman4@2402:b801:2844:2100::1)
05:56:24 Quit ideasman42 (Client Quit)
05:57:03 Join ideasman42 [0] (~ideasman4@2402:b801:2844:2100::1)
05:58:22ideasman42Hi, happy rockbox user here... checking IRC works
06:00
06:02:21rudi_sIt does
06:03:56ideasman42\o/ - general question, is it worth reporting crashes, in recent builds for modern hardware (currently in beta)
06:04:53ideasman42(xduoo x3ii currently crashes often when unplugging USB- for example), otherwise works well
07:00
07:26:01***Saving seen data "./dancer.seen"
07:36:39 Quit ideasman42 (Quit: WeeChat 3.1)
07:59:24 Join eevan [0] (~eevan@user/eevan)
08:00
08:37:07 Join cockroach [0] (~blattodea@user/cockroach)
08:37:51 Quit eevan (Quit: WeeChat 3.2-dev)
08:38:11desowinbraewoods: depends what you mean by unique, it is pretty much completely different hw than the other sansas
08:38:40desowinthe OF bootloader and RSA signature loading bug is pretty unique
08:39:09desowinanyway, the USB driver is up at https://gerrit.rockbox.org/r/c/rockbox/+/3454
08:39:38desowinhowever, I think this code is so big that I would say it is unreviewable
08:40:48desowinI have to figure out what upsets USB stack about the memory, when I unplug the bootloader loads app just fine from uSD card
08:42:12desowinI don't think it is the USB driver, but something in the Sansa Connect sdmmc code that I wrote nearly 10 years ago...
08:42:25desowinbut that's topic for another session...
08:47:40speachyideasman42: bugs are always worth reporting, especially if they are reproducable
08:55:10speachyideasman42: but in the case of the x3ii specifically, it's important to make sure you are using the most recent "bootloader" as that is what actually handles most of the USB hotplugging stuff.
09:00
09:08:23 Quit speachy (Quit: WeeChat 3.0.1)
09:26:03***Saving seen data "./dancer.seen"
10:00
10:19:48 Quit Natch (Remote host closed the connection)
10:28:45 Join Natch [0] (~natch@c-e070e255.014-297-73746f25.bbcust.telenor.se)
11:00
11:26:06***Saving seen data "./dancer.seen"
11:43:02 Quit johnb2 (Quit: Nettalk6 - www.ntalk.de)
11:59:28 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
12:00
12:23:07 Join speachy [0] (~speachy@209.2.65.77)
12:23:07Mode"#rockbox +v speachy" by ChanServ (ChanServ@services.libera.chat)
13:00
13:01:47 Join eevan [0] (~eevan@user/eevan)
13:05:06 Quit eevan (Quit: WeeChat 3.2-dev)
13:26:10***Saving seen data "./dancer.seen"
13:44:14 Join dconrad [0] (~dconrad@208.38.228.17)
13:56:35desowinhttps://gerrit.rockbox.org/r/c/rockbox/+/3455 this is something that can be reviewed :-)
13:57:35desowinbut unfortunately that's not making the USB MSC working yet, but fixes the strange screen displaying garbage due to reading data to the framebuffer...
14:00
14:12:08dconradamachronic: I can't get jztool to build, it can't find libusb - which I have installed and linked with homebrew, but I don't understand gcc/make well enough to know how to point it in the right direction
14:38:07 Join FroggestSpirit [0] (~FroggestS@d192-24-159-129.try.wideopenwest.com)
14:38:52FroggestSpiritI'm using the latest dev build of rockbox on my fii0 m3k, but the gbs player isn't working. it's silent
14:41:09rb-bluebotBuild Server message: New build round started. Revision 77603c344d, 297 builds, 10 clients.
14:41:54speachydesowin: do you want to keep poking at the tnetv105 driver or commit what you have so far in #3534?
14:43:25speachyI agree it's largely unreviewable except in the mechanical stylistic sense.
14:46:34FroggestSpiritIt seems to be the case with a fresh build too. The other libgme stuff still works like NES, SNES, Genesis
14:49:56rb-bluebotBuild Server message: Build round completed after 527 seconds.
14:49:57rb-bluebotBuild Server message: Revision 77603c344d result: All green
14:53:28speachyyay, a green board at last.
14:53:32 Join ats [0] (~ats@cartman.offog.org)
14:55:30 Quit ats_ (Ping timeout: 264 seconds)
14:57:10 Quit FroggestSpirit (Quit: Connection closed)
15:00
15:26:13***Saving seen data "./dancer.seen"
15:31:11 Join amachronic [0] (~amachroni@user/amachronic)
15:37:20amachronicdconrad: of course it just /had/ to not work... I can help if you can tell me what error you're getting, and where homebrew has dumped libusb libs.
15:37:36dconradyeah, give me a sec
15:39:41amachronicalso worth double checking, did you get the right libusb? You need libusb-1.0, sometimes it's called libusb1 and 'libusb' points to some old version.
15:40:06dconradhttps://pastebin.com/EzjBfeWV
15:40:16dconradlet me check that
15:40:59dconradlooks like I have libusb 1.0.24
15:41:20dconradlocated in /usr/local/Cellar/libusb/1.0.24
15:41:47amachronicheaders symlinked to /usr/local/include?
15:42:02amachronicI think I know the problem libusb.h is inside a subdir libusb-1.0.
15:42:59dconradyeah, I think? /usr/local/include/libusb-1.0 points to ../Cellar/libusb/1.0.24/include/libub-1.0
15:43:46amachronicok just edit the makefile and under the OSX section change -I/usr/local/include to -I/usr/local/include/libusb-1.0.
15:43:59dconradok, one sec
15:44:08amachronicthen we can see what the next error is ;)
15:44:17dconradhaha
15:44:19dconradyeah
15:45:17dconradis that the only thing we need under /usr/local/include then?
15:45:36amachronicyeah there's no other external libraries.
15:46:17dconradwell it built, I just got a warning
15:46:19dconradld: warning: directory not found for option '-L/opt/homebrew/lib'
15:46:34dconradactually, let me pastebin my whole result just for fun
15:46:36 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
15:46:50amachronicokay harmless, I had just read this was where homebrew puts things on ARM macs so I was trying to cover more cases.
15:47:17dconradhttps://pastebin.com/LFYx0SvQ
15:47:42dconradoh, yeah for full disclosure this is a 2015 macbook
15:48:04amachronicI'll have to find out how to test the correct arch and ifdef the right path I guess...
15:48:52amachronicnow, the real question is does it work to load up the bootloader? if you don't mind?
15:49:24dconradyeah, so I suppose I should put the original back on so it boots to the OF?
15:49:33dconradotherwise how will we know it changed
15:49:49amachronicnah it's unnecessary. Just plug it into USB boot mode
15:50:01dconradalright
15:50:10amachronicjztool doesn't actually do anything on the host end anymore, except load the bootloader
15:50:18amachronicso if it loads then it works
15:50:25dconradoh, cool
15:52:17dconradyeah, I got into the bootloader
15:52:22dconradno sudo needed either
15:52:35amachronicthank you!!! it all works then.
15:52:37dconradjust did ./jztool fiiom3k load bootloader.m3k
15:52:48amachronicthat was the plan :)
15:53:11dconradnow, is it statically linked or dynamically linked? do you want the executable?
15:54:25amachronicif you are able to generate a static binary speachy can upload it if you make it available somewhere
15:54:43dconradI figured static/dynamic is taken care of in the makefile...
15:54:53amachronicunfortunately you will need to edit the ldflags in order to make it output a static binary
15:55:03dconradI see, is it easy?
15:55:17amachronicit should just be a matter of adding -static
15:55:30dconradhmm, let me try
15:58:45dconradah, I'm not seeing any ldflags line in the makefile?
15:59:01dconradldopts
15:59:03dconrad?
15:59:03amachronicsorry it's called LDOPTS
15:59:05dconradah
15:59:45amachronicyeah not sure why the difference with rockbox, usually it's LDFLAGS :/
15:59:59dconradprobably an out of date version
16:00
16:00:19speachyI see both routinely
16:00:22dconradclang: error: linker command failed with exit code 1 (use -v to see invocation)
16:00:22dconradmake: *** [jztool] Error 1
16:00:47amachronicyou're probably missing static libs. I guess homebrew didn't build any.
16:01:01dconradhm, yeah that's probably what this is: ld: library not found for -lcrt0.o
16:03:03amachronicthat's C runtime stuff though, so it's probably an annoying apple toolchain thing. Unfortunately I know nothing of those matters so I'm not sure how to resolve it.
16:05:07dconradit would appear statically building executables isn't supported on macos? which is confusing, because I thought all apps for macos were statically linked
16:05:54dconradmaybe I'm thinking of .app files, which... are they dynamically linked internally? I dunno
16:06:28dconradso my understanding is, if I give this exe to somebody else, they need to have libusb in exactly the same place for it to work, right?
16:06:34amachronicI see. well, if the ABIs are sufficiently stable then dynamic linking against system libraries is not a problem but it is preferable to have a statically linked copy of libusb.
16:08:25dconradwell, as long as homebrew linked it right, I think it's pretty version-agnostic
16:08:42dconradproblem is, I don't know what macports does, and some people use one, some people use the other...
16:08:55dconradthough I think homebrew might be more common nowadays
16:12:28amachronicwell any path difference is moot if the ELF headers just have the soname and the dynamic linker searches for it (akin to LD_LIBRARY_PATH on linux)
16:12:54dconradso you figure it would be fine?
16:14:03amachronicI couldn't say, there are perhaps other issues on mac anyway, and I'm not sure it's worth uploading a copy unless we know it's going to work for most people.
16:14:26amachronicit isn't super user friendly to say 'do homebrew libusb' first
16:14:46amachronicbut it isn't user friendly to say 'go use Zadig install this driver' either, so...
16:15:34dconradyeah... I think I might be able to statically link against libusb while leaving everything else dynamic
16:15:47dconradwell, I'll try anyway
16:16:27amachronictbh I haven't had to do anything like that for a while but I think you would just pass -l/full/path/to/libusb.a
16:19:22dconradinstead of -lusb-1.0?
16:19:56amachronicyes althoguh I can't test my hypothesis on linux at this moment.
16:19:59amachronic*although
16:20:45amachronic(as libusb cannot be a static library unless you compile a whole new copy yourself thanks to udev)
16:21:30 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
16:21:32dconradso it's just giving me "ld: library not found for -l/usr/local/Cellar/libusb/1.0.24/lib/libusb-1.0.a"
16:22:47dconradwith "-l/usr/local/Cellar/libusb/1.0.24/lib/libusb-1.0.a -L/usr/local/Cellar" in ldopts
16:23:19dconradeh I have no idea what I'm doing haha
16:23:28amachronicwell lemme try spinning up a static libusb on linux and try it myself I'm not sure what options to use here either.
16:28:16 Quit tomato (Ping timeout: 245 seconds)
16:31:22 Join tomato [0] (~tomato@user/tomato)
16:31:42amachronictry getting rid of the -l option and pass libusb.a directly.
16:31:56amachronicit seemed to work for me
16:33:44dconradon the ldopts line?
16:34:31amachronicyep. you might have to add some system libs as well if the final link fails
16:36:21 Join tomato6 [0] (~tomato@user/tomato)
16:36:53dconradit just can't find it, apparently
16:37:10dconradclang: error: no such file or directory: 'libusb-1.0.a'
16:37:24amachronicwell I assume it exists of course?? did you give the full path?
16:37:31 Quit tomato (Ping timeout: 264 seconds)
16:37:31 Nick tomato6 is now known as tomato (~tomato@user/tomato)
16:37:41dconradhm, I didn't give the full path without -l, I'll try that
16:38:22dconradoof, that found /something/
16:39:22dconradUndefined symbols for architecture x86_64: ... and then a whole list of things from libusb-1.0.a
16:39:35dconrad...maybe this isn't worth the effort
16:40:03amachronicyep if you recognize the symbols you need to link in the correct system libs because the libusb.so isn't transitively pulling them in anymore.
16:40:53amachronicif you pastebin it I can try to look them up.
16:41:59dconradif you want to, it's here: https://pastebin.com/WTJEYnzZ , but... that's a lot
16:43:22amachronictry adding -framework IOKit -framework CoreFoundation to ldopts
16:44:09dconrad...
16:44:13dconradwell, it built
16:44:23dconradyou must have some pretty good google-fu
16:45:01amachronicI just pasted in _CFDataGetBytes and got an identical problem...
16:45:03amachroniclol
16:45:31dconrad... and it booted!
16:46:14dconradI don't know how to test this though, if it's enough to maybe uninstall libusb through homebrew?
16:46:24dconrador find somebody that doesn't have homebrew at all
16:46:36amachronicyeah just uninstalling libusb is a quick'n'dirty way to assure yourself
16:47:00amachronicbut you should be able to run ldd jztool (or whatever the mac equivalent is) and see what shared libs are being used.
16:49:16dconradwell I uninstalled all versions of libusb and it still worked, no libusb-1.0 in /usr/local/include anymore
16:52:39dconradyeah, I think it worked as intended - https://pastebin.com/eg4vTsRU
16:53:06amachronicit _should_ be okay now, barring any ABI issues in the system libs.
16:53:21amachronicso I assme that compatibility version stuff means it will work on older versions of macos?
16:53:31dconrad... I would guess?
16:54:55dconradhmm, my roommate has a couple really early osx machines, maybe I could try it on those for fun
16:55:23dconradit'll have to wait a few days tho
16:56:31amachronicah well, it's probably safe. we can just hope apple put some effort into proper ABI versioning for their core libs. naively I think the fact it's reported at all would suggest so.
16:57:01amachronicbut if you would like to wait and test it then by all means do so. I'm in no rush.
16:57:27dconradwell, I might as well get you the executable and we can see if anybody complains :-P
16:59:04amachronicalright then, feel free to upload it.
17:00
17:06:02dconraduh... is that through wiki?
17:07:17amachronicmeh just upload it anywhere, dropbox or google drive etc., speachy has to grab it and put it on the download site.
17:07:36dconradok, got it
17:08:29dconradit's at dconrad.neocities.org/assorted_files/jztool">https://dconrad.neocities.org/assorted_files/jztool
17:09:24amachronicawesome, thanks a bunch.
17:09:38dconradno problem, sorry for being clueless haha
17:09:41amachronicspeachy: ^^^ want to upload that to the m3k bootloader files?
17:11:00amachronicand don't worry about it :) you've been a real help.
17:12:03dconradoh, I tried it on my other roommate's m1 mac mini, works just fine :-D
17:12:18dconrad(I totally forgot about it until just now haha)
17:12:44dconraddefinitely doesn't have libusb
17:13:28amachronicm1 as in ARM? guess they had to give into an entire x86_64 emulator haha.
17:14:09dconradyeah there's some kind of really slick compatibility layer
17:14:15dconradbut 64 bit only
17:14:39dconradhonestly, their processor is super impressive
17:15:19amachronicyou need all that perf to spend more cycles emulating x86... lol.
17:16:14dconradI think it might be emulated in hardware? but anyway yeah I think it's pretty seamless
17:26:16***Saving seen data "./dancer.seen"
17:31:37braewoodsdconrad: i'm not impressed but then it's probably because of how limited it is.
17:39:54 Quit amachronic (Ping timeout: 264 seconds)
17:48:20speachydconrad: just to be clear, this is a MacOS build? x86_64?
17:49:39speachyit's up, btw.
17:49:44dconradyeah,
17:50:20dconraduhh built on 10.15.7 (Catalina)
17:50:50dconradworks on brand new m1 macs, unknown how old you can go
17:51:51dconradx86_64, yep
18:00
18:15:33 Join K4os_ [0] (~K@ip5f5af5c5.dynamic.kabel-deutschland.de)
18:21:50braewoodswhat would impress me is if someone creates an ARM motherboard that is similar to existing PC boards
18:22:10braewoodsin terms of raw capabilities
18:23:03K4os_apple?
18:23:23braewoodsK4os_: does it offer expansion slots, socketed memory, etc?
18:23:41K4os_via firewire - yes
18:23:52braewoodsno pcie then?
18:23:58K4os_afaik no
18:24:07braewoodsok then not really comparable :|
18:24:24K4os_isnt firewire able to do pcie stuff too ?
18:24:30K4os_like external GPUs
18:24:39braewoodsfirewire isn't in modern macs.
18:24:45braewoodsit was replaced with something else.
18:24:47K4os_well usb c then
18:24:53K4os_lightning
18:24:57K4os_idhnfc
18:25:04braewoodsand i dunno about that
18:25:28K4os_anyway, id like some serious ARM/RISC hardware,too - if it would be open
18:25:47braewoodsthe closest thing is probably pine microsystems stuff
18:25:50braewoodsfor now
18:25:51K4os_but right now i would really prefer to have rockbox running on my new mp3 player =P
18:26:38braewoodsthe downside of ARM is there's no standard architecture unlike x86
18:27:34braewoodsso you end up with stuff revolving around a vendor's implementation of it rather than a standard reference platform
18:28:01braewoodsthis is a byproduct of the PC clone era
18:28:37braewoodsit's funny, it's like we've moved away from clones and that level of hardware interoperability
19:00
19:07:33 Quit lebellium (Quit: Leaving)
19:19:01speachythe PC really was the exception. well-documented expansion hooks, the BIOS provided a unversal hardware-access API, and later busses were probe-able in a safe, non-destuctive, self-describing manner.
19:20:41speachyMacs never had to be interoperable with anything else. Apple controlled everything end-to-end so there never was any question of compatibility within its own ecosystem.
19:21:54speachymodern ARM (at least in the server space) requires ACPI to hide all of that, with standard boot interfaces too.
19:23:07speachybut on the linux side of things, standardizing on the devicetree finally allowed for generic Linux images. still had unique bootloaders, but that's no different than the average PC either.
19:26:20***Saving seen data "./dancer.seen"
20:00
20:27:54 Quit dconrad (Remote host closed the connection)
21:00
21:26:24***Saving seen data "./dancer.seen"
21:42:53speachyhmm, I should add a set of toolchain builds to the build farm..
22:00
22:14:38 Quit cockroach (Quit: leaving)
22:20:46 Quit K4os_ (Ping timeout: 245 seconds)
23:00
23:26:26***Saving seen data "./dancer.seen"
23:32:03 Quit kadoban (Read error: Connection reset by peer)
23:32:04 Quit blbro[m] (Read error: Connection reset by peer)
23:34:00 Join kadoban [0] (~kadoban@user/kadoban)
23:39:19 Join blbro[m] [0] (~blbrostra@2001:470:69fc:105::8f7)

Previous day | Next day