#milkymist IRC log for Wednesday, 2012-02-08

lekernelpress the three buttons at the same time00:05
dvdkok, that works but is a little tricky to do00:08
dvdklekernel: that doesn't seem to work if my software screws up (all interrupts masked, BTW)00:19
dvdklekernel: oh, it does work, i just need to push harder00:20
kyakdvdk: have you moved to a different time zone or just can't sleep? :)00:45
dvdkkyak: i've moved to a different work schedule :)00:52
dvdkkyak: you here on mm?  own one?00:52
kyakyeah, recently got one :)00:53
kyakdvdk: btw, have you seen this error: http://dpaste.com/699279/00:55
kyaktrying to follow instructions in milkymist/scripts/README.html00:55
dvdkkyak: no, but then I haven't tried to run filickernoise compilation yet00:55
dvdk(only installed the rtems toolchain)00:55
kyakwhat are you playing with so far?00:56
dvdkkyak: making gforth-ec run.  see here https://github.com/earl/gforth-mirror/tree/master/arch/lm32 (not 100% up to date, still buggy/broken)00:57
dvdkthis is very minimalistic.  doesn't need binutils, nor lm32-gcc to compile :)00:57
kyakhm, what does it need then to compile? :)00:59
kyakare you trying to make gforth-ec run instead of flickernoise or together with flickernoise?01:00
dvdkkyak: instead of it.  just to get a low-level debug console for hacking with (new) SoC components01:01
kyakdvdk: cool! how do you replace it?01:10
dvdkkyak: currently just booting the gforth-ec image from the bios (press F8 on serial term during bootup, type serialboot, using 'flterm' tool from milkymist.git01:11
Action: fpgaminer should get an NES emulator running on Milkymist...01:11
wolfspraulkyak: dvdk both of you here - such an honor!01:16
wolfspraulwelcome and good morning from Beijing ;-)01:16
kyakdvdk: i see :)01:21
kyakxiangfu: hi!01:22
xiangfukyak, Hi.01:22
xiangfukyak, great to see you here #milkymist :)01:23
xiangfukyak, Hi. just check the log.01:26
kyakxiangfu:  thanks :)01:26
xiangfukyak, the compile error is because the recently upstream break flickernoise. :(01:26
xiangfukyak, revert to 'f80b3a3d825110b5d8826f72db3fa47a6d71b66a'01:27
xiangfugit co -b  fn f80b3a3d825110b5d8826f72db3fa47a6d71b66a :)01:28
kyakah, lucky me :)01:28
xiangfukyak, you can use tftp to replace flickernoise with other images. like the linux01:30
xiangfukyak, when milkymist boot. press 'ESC', then it will goto bios console.01:31
xiangfuthe command 'netboot' will try to load 'boot.bin' from
kristianpauldvdk: no need reset button at least your urjtag really wont work01:32
kristianpaulwow gforth-ec !! tell us more about it01:33
kristianpaulwhat can you do? :)01:33
kristianpauldvdk: http://projects.qi-hardware.com/index.php/p/wernermisc/source/tree/master/m1/jtag-boot01:34
kristianpauldvdk: i basically use that when not at home and want to load some bitstream when i'm getting boring at work :/01:34
dvdk`kristianpaul: don't understand, i can reboot the mm via urjtag?  how?  which command?01:35
kristianpauldvdk`: make boot01:35
dvdk`which makefiles01:35
xiangfudvdk`, http://pastebin.com/e18P5Da6  I have a m1reboot.sh under ~/bin01:36
xiangfukristianpaul, things get easy now. :)01:36
kristianpaullets see01:36
kristianpaulxiangfu: havent updated milkymist repo, you mean that ~/bin is there?01:37
dvdk`xiangfu: very, very nice.  thanks a lot :)))01:38
dvdk`should put that into the wiki.01:38
xiangfukristianpaul, I have m1reboot.sh m1bios.sh and m1usb.sh01:38
dvdk`xiangfu: you just increased the life expectance of these buttons 100-fold.01:38
xiangfum1bios.sh : for compile and reflash bios.01:38
xiangfum1usb.sh : when you modify the softusb-input. it's for compile bios. reflash bios. update rtems. make rtems and re-compile flickernoise01:39
kristianpaulwhat happened with the makefiles?..01:39
xiangfukristianpaul, it's using the makefile. just make 3 commands to 1. :)01:40
xiangfudvdk`, wiki. yes.01:40
xiangfudvdk`, one question about your PATCH email.01:40
dvdk`the one from today or the earlier one?01:41
xiangfuyour email is not plain text here.01:43
xiangfuthe earlier one already applied. thanks.01:44
dvdk`xiangfu: you mean all email, or just the one from today?01:44
xiangfuafter I save as a .eml file. it give me: http://pastebin.com/VztmEugT01:44
xiangfuall those two patches under milkymist mailing list.01:44
xiangfuthe earlier one I have to manually edit a little before apply.01:44
dvdk`oh I hate it.  i send via emacs (gnus), maybe that's a bad idea.   next time i'll just append the patch as an attachment01:45
kyakxiangfu: thanks! btw, step 6 in README.html should have "milkymist" instead of "milkymist.git"01:46
dvdk`xiangfu: or your mail-reader is just not smart enough to un-encode the mail (quoted-printable or something)01:46
xiangfudvdk`, try git send-email :)01:46
dvdk`xiangfu: no way :)  via smtp.  well will have a look.01:47
xiangfudvdk`, git send-email support smtp.01:48
wpwrakxiangfu: btw, regarding your report patches: you're a bit ahead of me :) the various usb errors you're seeing are "normal". also other devices produce them. so it's hard to tell whether any regression your patches may or may not produce would come from disturbing whatever is wrong in the lower level of usb or whether there's a real bug in them.01:48
wpwrakxiangfu: i also suggested to sebastien to give you commit access to milkymist.git, so that you can send your patches in a topic branch. then it'll be easy to use them in builds, and also to merge them once the stack is a bit more stable.01:50
xiangfuwpwrak, thanks.01:50
kyakxiangfu: hmm, can't find the commit f80b3a3d825110b5d8826f72db3fa47a6d71b66a.. Which repo is it?01:51
xiangfukyak, the rtems repo.01:52
GitHub60[scripts] xiangfu pushed 1 new commit to master: http://git.io/mTTWtA01:52
GitHub60[scripts/master] update readme with new folder name - Xiangfu Liu01:52
xiangfukyak, fixed ^ :)01:52
kyakthanks :)01:52
kyakmake[7]: *** No rule to make target `/opt/milkymist/rtems/c/src/../../cpukit/score/src/heapiterate.c', needed by `src/libscore_a-heapiterate.o'.  Stop.01:54
kyaknow it's this error, after i checked out to that commit01:55
xiangfuwpwrak, I will create a branch name softusb-input. then apply those patches to it. is that ok?01:55
xiangfutotally is 24.01:59
wpwraklovely :)01:59
xiangfukyak, under compile-flickernoise01:59
kyakthis whole thing is kind of tricky.. some directories in opt, some in milkymist-scripts.. :)01:59
wpwrakthe HID designers probably got some extra bonus for making things complicated01:59
xiangfuthe factory better follow the boot protocol.02:00
wpwraki had some hopes for the boot protocol. but if we lose half of a device's functionality with it, then it's not good enough02:01
xiangfuwpwrak, is that last patch works fine with riitek rf keyboard?02:04
wpwraki didn't try yet02:04
xiangfudvdk`, oh. just check the email it's 'Content-Transfer-Encoding: quoted-printable'02:05
dvdk`but doesn't git undedrstand that encoding?02:05
xiangfudvdk`, I don't know so.02:06
xiangfuI don't think so. since I already test it with git am :)02:06
GitHub91[milkymist] xiangfu pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/0a5dc0899af0ed6abd3f8dd7c88aa89f4bc007ab02:11
GitHub91[milkymist/master] fix 'make install' target - David Kuehling02:11
xiangfuwpwrak, oh. I got the DX2, LV3 and icon from wolfgang yesterday. :)02:11
xiangfudvdk`, applied. :)02:12
wpwrakxiangfu: now you can have some REAL fun ;-)02:13
wpwraki hope wolfgang has some left to show to jon. our traveling salesman needs to know what M1 can do, too :)02:13
xiangfuwhen I plug-in the LV3 is give : http://pastebin.com/rnSzC3Gv02:16
wpwrakxiangfu: all that matters is "Found MIDI" :-)02:19
xiangfuwpwrak, ok. :)02:19
wpwrakxiangfu: that is, unless you want to debug the lower-layer problem. i suspect it's some timing issue.02:19
wpwrakmaybe some weird inter-packet timing requirement. or we're too slow to retrieve data at some point.02:21
kristianpaulwhere is that built-in logic analizer when needed..02:23
wpwrakyes, indeed :)02:24
kyakxiangfu: the download location for libpng-1.5.6 on SF has changed.. it is project/libpng/libpng15/older-releases/$(LIBPNG_VERSION)/$(LIBPNG)" now (line 259 of compile-flickernoise/Makefile)02:25
wpwrakkristianpaul: right now, you still need a scope with at least 50 MHz bandwidth, deep memory, and support by my USB protocol decoder (that it, unless you happen to have a good USB protocol analyzer, in which case i shall envy you)02:27
kristianpauldont think so..02:30
kristianpauland  deep memory sounds expensive02:30
kyak/bin/sh: lemon: command not found - the hell is that..02:30
GitHub43[scripts] xiangfu pushed 1 new commit to master: http://git.io/af73Bw02:30
GitHub43[scripts/master] update libpng download URL. thanks kyak - Xiangfu Liu02:30
kristianpaulkyak: what are you compiling?02:31
kristianpaulahh whole nightmare by hand ;)02:31
kyakit;s almost done02:32
xiangfukyak,  apt-get install lemon02:32
xiangfukyak, needs ask Werner about lemon. :)02:32
kyakit's not in Arch's repos02:33
kyakThe LEMON LALR Parser Generator used in sqlite 1.6902:33
kyakit's in AUR.. let's hope this is what i need :)02:33
kristianpaulsuck less parser i was told, but seems wpwrak disagree a bit? :)02:37
wpwrakkyak: you'll also need re2c. sebastien tried very hard to avoid the more conventional (f)lex and yacc/bison :)02:38
wpwrakkristianpaul: it's a mixed experience. some things are nice. others not so.02:38
kyakwpwrak: i also need avr-gcc, apparently :)02:39
wpwrakkristianpaul: (deep memory) not terribly so. the rigol ds1000c/cd/d/e series devices all are suitable. (not the 1000b, though)02:40
wpwrakyes, you do :)02:40
kyakok, it's done02:40
kyaki had to install libftd2xx, urjtag, llvm, clang, avr-gcc, lemon and re2c overall :)02:42
kyaknow, is that .fbi file is the one i should tftp to mm?02:44
xiangfukyak, no. .fbi is for flash02:44
xiangfuthe /flickernoise.bin is for tftp02:45
xiangfurun 'make load' under flikernoise/src02:45
xiangfuit will compile .bin file and copy it to /var/lib/tftpboot/boot.bin02:46
kyakoh.. ok.. i have to switch back and forth between /opt/milkymist and milkymist-scripts02:47
xiangfukyak, in fact make -C compile-flickernoise is only needs once. other time you will always play with milkymist, rtems and flickernoise :)02:47
kyakwhy did i need to build that flickernoise.fbi?02:50
kristianpaulto use with urjtag02:51
wpwrakkyak: you can also flash it into NOR and then boot from there. slightly slower but you can be sure the code doesn't wander off when you reboot02:52
kristianpauland flash it02:52
kristianpaulto the NOR harcoded partition for flickernoise02:52
kyakah ok..02:52
kristianpaulactually is called APP02:54
wpwrakkyak: it's pretty easy with http://projects.qi-hardware.com/index.php/p/wernermisc/source/tree/master/m1/tools/m1nor02:54
kristianpaulindeed :)02:54
wpwrakkyak: m1nor flickernoise.fbi   that's all02:54
wpwrak(well, once you have jtag working. that can sometimes be a bit of trouble)02:54
kristianpaulyeah Fallenou reported some  trouble last time..02:55
kyaksounds easy, but undocumented :)02:55
kristianpaulall is recorded here :)02:55
wpwrakmaybe not perfectly up to date but close to what i use02:56
kyakstill i find it confusing that some files are under milkymist-scripts/compile-flickernoise and some are in /opt/milkymist/flickernoise02:56
kristianpaulmilkymist-scripts/compile-flickernoise should be just temporary or sources archives no more i guess? xiangfu ?02:58
wpwrakdo you even needs these ?02:58
xiangfukristianpaul, kyak it's like. after you get the .fbi file. that mean all toolchain and libs are installed in your system. then you have to jump to /opt/milkymist stuff to hack.02:59
kristianpaulbut yes now wiki is some behind milkymist.org i guess is time to have a better organized and good looking place? just move02:59
kyakxiangfu: do you mean that once i ran "make -C compile-flickernoise flickernoise.fbi" i can remove the milkymist-scripts/compile-flickernoise and continue in /opt/milkymist/flickernoise?02:59
kristianpauldoxygen ! (kidding :)02:59
kyakxiangfu: for example, libpng. This is a part of flickernoise, right? Where did it go?03:01
kyakthere are no traces of libpng under /opt/milkymist/flickernoise03:01
kristianpaulgo to rtems in /opt/rtems*03:02
kristianpauli remenber..03:02
xiangfukyak, yes. part of flickernoise. /opt/rtems-4.11/lm32-rtems4.11/milkymist/lib/libpng.a03:02
kristianpaulno no part of FN, rtems have the lib but seems this one is better updated?03:02
kristianpauli dont remenber whole reason03:02
kristianpaulor just lack it03:03
wpwrakkyak: i would recommend building things directly in the repositories, without trying to do a "world" build03:03
xiangfukyak, flickernoise is a big APP. that links all those libs and rtems to only one.03:03
kristianpaulBIG app :)03:03
wpwrakkyak: when developing, you'll want selective builds anyway03:03
kyakso does that .bin file contain both RTEMS and flickernoise?03:03
xiangfukyak, yes.03:04
kyakok, i got it03:04
xiangfurtems is not like Linux kernel. it more like libs.03:04
xiangfulink to target.03:04
kyakwpwrak: i understand what you recommend.. i just decided that it is easier to jump in with xiangfu's Makefiles03:04
kristianpaulnewlib latelly no?03:05
kyakso it's like a single application linked with a bunch of libs, right?03:05
kristianpaulflickernoise yes03:05
kristianpaulyeah !03:06
kristianpaulyou could get for example things like lua pretty easy running03:07
kyaki'd like to have busybox running on RTEMS :)03:07
kristianpaulbecause you already got the board support package from rtems03:07
kyakdoes it mean that i will have ONLY lua running?03:07
kristianpaulkyak: busybox is ansi c and no mmu, so i guess so you can build in against rtems not ulibc03:08
kristianpaulkyak: yes03:08
kristianpauljsut straight a lua itnerpreter :)03:08
kristianpauls/s/to a03:08
kyakyeah, this is all definitely different from Linux..03:08
kristianpaulsimpler (at least for me not used to hack linux)03:09
kyakbut it's a good opportunity for real-time applications03:09
kyakok, time to wake up and go to work.. Wonder when i'm gonna stop jet lagging03:10
xiangfukyak, there is a Linux port : https://github.com/milkymist/linux-milkymist03:10
kristianpauli like the dvdk idea porting gforth, compile over and over again is so boring..03:10
kristianpaulah yes03:10
kristianpauland run busybox, pretty cool :)03:10
kyakxiangfu: suddenly running Linux on it is not so interesting, have to explore RTEMS :)03:10
kristianpaulkyak: well there is ELF  wich i missed a bit..03:11
kristianpaulfrom Linux :)03:11
kristianpaulat least you can run a small c compiler on the M1 it self that will be COOL03:12
Action: kristianpaul looks for wpwrak opinion on the topic03:12
xiangfuwpwrak, is that ok I push those four hid patches to master ?  I think the code is ok. and do fix some bug.  :D03:12
kyakyou can run a c compiler on m1, but you can;t run the executable (unless you reboot into it) - right? :)03:12
wpwrakxiangfu: why not to the "hid" branch as we discussed ?03:16
wpwrakor is this something new ?03:16
xiangfujust want faster goto 'master' branch. :)03:19
kristianpaulyou'll need some tests first i guess :)03:20
xiangfuyes. let's do hid branch first. then other people can easy test.03:21
GitHub71[milkymist] xiangfu created hid (+4 new commits): https://github.com/milkymist/milkymist/compare/6587145^...b05e7e703:22
GitHub71[milkymist/hid] pass ep_status to ep process - Xiangfu Liu03:22
GitHub71[milkymist/hid] softusb-input: add GET_HID_REPORT_DESCRIPTOR support - Xiangfu Liu03:22
GitHub71[milkymist/hid] cleanup code style, remove space after if - Xiangfu Liu03:22
wpwrakyeah, we already have regressions between july and now. with too many changes on top of the shaky foundation, i'd worry about the thing getting out of control.03:23
wpwraknice, branch creation worked ! :)03:24
xiangfu(i'd worry about the thing getting out of control.) agree03:28
xiangfuwpwrak, I tried midi2osc yesterday. when I plug-in the LV3 to my laptop. then start ./midi2osc -d then nothing output.03:31
xiangfunothing output when I press any button on LV3.03:31
xiangfuthe code stop at 'while (snd_seq_event_input(midi, &ev)) {' ?03:32
xiangfuI want find out the channel and control on those devices.03:38
wpwrakdid you connect it with qjackctl ?03:45
wpwrak(connect it) the LV3 to midi2osc03:45
wpwrakor, if you connect multiple devices, all of them to midi2os03:46
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120208-0301/03:48
wpwrakand to remap the LV3 controls to things FN can handle, you can use  ./midi2osc `sed 's/#.*//' LV3`
wpwrak... or use the direct-midi branch :)03:49
xiangfuwpwrak, ah. I have to connect them together. then it will output the values03:49
xiangfuif there a way that no needs qjackctl? just output those channel and control values messages?03:50
wpwrakthere probably is ... if you find it, please post how it's done :)03:51
xiangfuconnect them means send the LV3 output to midi2osc. midi2osc is not direct open the midi device.03:53
wpwrakin any case, once direct-midi is merged, you'll need midi2osc only if you have more MIDI devices than USB ports. FN won't need remapping anymore.03:53
cladamwwpwrak, if you are available, could you review my dvi email? but take your time. thanks !03:54
xiangfuwpwrak, I want make sure the channel and controls. the encoder change values real small.03:55
wpwrak(remapping) that is ... it still might. e.g., if two devices use the same addresses.03:55
wpwrakxiangfu: i didn't understand that03:56
xiangfufor example. I turn a whole around of 'Mater Volume'. it output: http://pastebin.com/wvmEjAaY03:58
xiangfuit's 1, 2, 4, stuff. I think it should be 0 ~ 12703:59
wpwrakah ;-)03:59
wpwrakthe knobs send "acceleration" values. it's sort of relative values03:59
wpwrakhere's how i process such things in the direct-midi branch: https://github.com/milkymist/flickernoise/blob/direct-midi/src/renderer/stimuli.c#L4004:01
xiangfuah. got it.04:01
wpwrakyou could also do it in the patch, but it would be messy04:02
wpwrakcladamw: that atlys boards seems to create more questions than it provides answers :)04:05
cladamwwpwrak, yes, but from Chinese links, seems that we are close to dvi good design. but let's be careful04:06
cladamwwpwrak, i may oder a DVI-I monitor later to measure its pin status, but it's obviously from links. ;-)04:07
wpwrakbah. quits the second i send the reply :(04:31
cladamwwpwrak, thanks for reminds on HPD pin to fpga 5V tolerance, i forgot it. ;-)05:41
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120208-0458/05:46
wpwrakcladamwa: glad that i could contribute something ;-)08:34
cladamwawpwrak, u re modest. as usual though. :-)08:37
cladamwawpwrak, i'm checking joerg's great feedback on audio. any idea from you too ?08:39
cladamwawpwrak, for all unused inputs, from your last email, i added couple TPs for keeping the DC block footprints08:44
cladamwa  (C4 through C8), so that one could solder something external08:44
cladamwa  (just in case). from Joerg's idea on two pins with jumper, it's also not bad but occupying pcb space.08:44
wpwrak(audio pins) i was just waiting for that audio ground to return ;-)08:54
wpwrakwell, it's just one more pin, so ...08:54
cladamwamm? Joerg's idea on J23 i think it's pretty good like you want.08:55
cladamwa(J1) i think too.08:56
cladamwa(100R) C19*R17 <-100R-> J2:2,3 and C21*R18 <-100R-> J2:4,5; this i don't understand .08:59
wpwrak(j23) yes. i'm only surprised he didn't suggest it earlier. when we discussed it, i was gently prodding in this direction, but he seemed happy with just the floating ground.09:04
lekernel__wpwrak: what are the remaining regressions?09:04
lekernellet's push out webupdate images soon09:04
wpwraklekernel: some devices that worked in july not working now, for some unknown reason.09:06
lekernelok, just USB stuff?09:07
wpwraklekernel: my plan is to set up a proper record for my mouse collection and then go after those timeouts09:07
wpwrakyes, just USB09:07
wpwrak(mouse collection) i have something like ten different mice now. so that should be a useful base for regression checking. needs careful record-keeping, though, or it'll end up in a big ghost chase09:08
wpwrak(ghost chase) trying to debug imagined regressions, and also overlooking real issues. i'd also imagine that not all problems happen all of the time.09:09
wpwrakcladamwa: he wants to insert 100 Ohm series resistors, in case something nasty is connected09:09
cladamwalekernel, wpwrak, so 100 Ohm series are okay to you?09:11
wpwrakwhat's typical line impedance again ? 1 kOhm ? 10 kOhm ?09:12
lekernelphew, imo this whole audio improvement thing is overkill. I'm out of it :)09:13
cladamwalekernel, overkill ? ;-)09:14
lekernelyes, overdoing things, fixing something that works, etc.09:14
wpwraki guess as long as all you use is treb, mid, and bass level, then it's indeed overkill ;-)09:15
cladamwawpwrak, each of LINEOUT has 10 KOhm to AUDIO_AGND.  did you try to say this ?09:16
lekernelwpwrak: by the way, we already have a 47 ohm resistor in series with the DCC 5V09:16
lekernelif you want to make that thing more resistant, just use a resistor that can dissipate ~500mW permanently... the diode won't protect against short circuits09:17
wpwrakcladamwa: hmm yes, that fits. okay, so adding 100 Ohm just increases overall impedance by 1%. negligible.09:17
wpwraklekernel: heh, or pick a diode with a large Vf @ 100 mA ;-)09:19
wpwrakbut yes, it should be short-circuit proof09:20
wpwrakno clamp diodes in the FPGA ? wow09:21
wpwrakthere a "Maximum current through pin using PCI I/O standard when forward biasing the clamp diode.", so that's something else ?09:24
wpwrak(in DS162)09:25
lekerneland the clamp diodes aren't there by default (eg you can apply voltage to the I/Os with VCCIO turned off, and no current will flow)09:43
lekernelwe can also use a voltage divider, which solves pull-down and level conversion at the same time... but after we've had so many solder and misplaced component issues, I'm not sure if it's safe09:48
wpwrak(clamps) thanks ! one more thing learned about those tricky fpgas ...09:51
wpwrakhow would a voltage divider be less safe than, say, a FET ?09:52
wpwraki'm sure you can have some hellish smt mishaps with a fet as well :)09:52
lekernelif the bottom resistor isn't there or has the wrong value, the FPGA can get too much voltage09:52
lekernelif the FET isn't there, the FPGA simply doesn't get any signal09:52
wpwrakhmm okay, absence / no connectivity is more likely than other sorts of problems, agreed09:53
cladamwa(J1) wpwrak, have you seen joerg's idea? is okay to you? "rationale: connect unused input to GND -> less noise"09:57
cladamwawpwrak, lekernel do you mean that we could use 2N7002 like Q1 to be as level shifter(5V -> 3.3V) ?10:05
cladamwalekernel, got it. ;-)10:06
wpwrak(J1) sounds good to me, yes. of course, if you want to detect plug insertion, things may change again. not sure how to do proper plug detection, though.10:15
wpwrak(plug detection) in my LED function description, i carefully avoided mentioning plug detection for almost anything ;-)10:16
cladamwawpwrak, he. i noticed you were gently avoid. but it's okay. we do this later for overview. ;-)10:27
cladamwa(J1 plug detection) I'd have some phone schematic from China, i'll see how they do such feature. ;-)10:35
fpgaminerHmmm, trying to compile milkymist gives me a compilation error "unknown field 'fpathconf_h' specified in initializer" when it tries to compile rtems_yaffs.c10:38
wolfspraulgently avoid ;-)10:41
cladamwa(M1R4 audio) http://downloads.qi-hardware.com/people/adam/m1/tmp/m1r4/Audio_20120208.pdf10:51
cladamwai didn't change those DC block capacitor's TPs which joerg suggested that we can use 2 pin jumper. since i hope there's enough space to do such but ... ;-)10:54
wpwrakah, nice. rtems has a live repository again13:48
Fallenou_yep and they have a gitweb14:06
GitHub178[milkymist-ng] sbourdeauducq pushed 2 new commits to master: http://git.io/hySOvA14:14
GitHub178[milkymist-ng/master] tools: remove bin2hex - Sebastien Bourdeauducq14:14
GitHub178[milkymist-ng/master] tools: use install and /usr/local (as suggested by David Kuehling) - Sebastien Bourdeauducq14:14
wpwraklekernel: when communicating with the user, MIDI channels are commonly one-based. in FN, they're 0-based. what would you think of changing that ?15:25
lekernelyes, then change it15:36
lekernelhmm, there seems to be some C++ working with clang/lm32 too :) (not that I like C++, but it's nice to see that with LLVM it seems other languages can be made to work without much fuss...)15:38
lekernellol, why the hell are the new/delete operators internally named "_Znwm/_ZdlPv" in C++? couldn't they think of even less meaningful names?15:42
wpwrakthey should have avoided the common _Z prefix. makes it too easy to remember.15:42
lekernelobjective-c appears functional, too15:49
lekernelI don't know it much but at least they do not have insane name mangling there :)15:51
kristianpaulgood catch17:27
wpwrakmmh ?17:31
kristianpaulgnustep similarity with osx?17:31
wpwrakdunno. both look like guis :)17:32
kristianpaulthat too :)17:32
kristianpaul23321 paul      20   0  312m 280m  12m R 56.5 31.8   2:31.02 xst18:06
kristianpaul20638 paul      20   0  371m 250m  19m R 39.3 28.4   6:18.31 map18:06
--- Thu Feb 9 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!