Wednesday, 2015-09-23

*** tpb has joined #timvideos00:00
CarlFKum... still building...01:00
CarlFKUpdating file: atlys_hdmi2usb-hdmi2usbsoc-atlys.ncd with current fully routed design.01:01
CarlFK  Intermediate status: 765 unrouted;       REAL time: 1 hrs 56 mins 6 secs01:01
mithroCarlFK: I haven't figure out why - but it seems sometimes the build takes 2.5 hours - most of the time it takes 30 minutes...01:27
CarlFKmithro: huh. I guess this is one of those times.01:28
mithroCarlFK: if you can make it happen reliably, that would be really useful01:29
CarlFKmithro: huh. I am just running bootstrap.. with the little tweaks I submitted.  which I don't think are part of the build process01:32
mithroCarlFK: yeah - same thing travis does01:32
*** travis-ci has joined #timvideos02:12
travis-ci[timvideos/HDMI2USB-misoc-firmware/mithro.get-env-root-fix#169] (a7a0675): The build passed. (
*** travis-ci has left #timvideos02:12
*** travis-ci has joined #timvideos02:13
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#168] (df8c617): The build passed. (
*** travis-ci has left #timvideos02:13
CarlFKyay, I have a prompt again!  [FLTERM] Starting... HDMI2USB>02:28
CarlFK video_matrix connect pattern encoder; encoder on; mplayer tv:// -tv driver=v4l2:device=/dev/video1 and I have a pattern.  Whew.02:31
CarlFKFPGALink Command-Line Interface ... built with no wgets!03:00
CarlFKthere is however...         rm -rf argtable2-${AT2_VERSION};  tar zxf argtable2.tgz
tpbTitle: libargtable2/ at master · makestuff/libargtable2 · GitHub (at
CarlFKwhich I think needs to be cleaned up or tumbleweed will twitch03:02
CarlFKtumbleweed: do you have 5 min to give packaging guidance?03:43
CarlFKI am guessing I need to do some grunt work, but maybe there is a more elegant way03:44
tumbleweedCarlFK: yeah, sure04:27
CarlFKtumbleweed: yay!04:28
CarlFKI am trying to sort out the code and build process so this can be packaged
tpbTitle: makestuff/flcli · GitHub (at
CarlFKsec.. let me post the script needed to grab all the sources...04:28
tumbleweedso, you'd need everything in a tarball04:29
tumbleweedyou can't fetch stuff from the net at build time04:29
tumbleweed(or multiple tarballs)04:29
CarlFKthere is however...         rm -rf argtable2-${AT2_VERSION};  tar zxf argtable2.tgz
tpbTitle: libargtable2/ at master · makestuff/libargtable2 · GitHub (at
CarlFKthat and the patch stuff below it..04:29
CarlFKmake my eyes hurt.04:30
tumbleweedpatches in debian packages are easy04:30
tumbleweedlibargtable2 is already patched, though04:31
CarlFKtumbleweed: with the above patch?04:32
CarlFKI went here and gave up trying to figure it out:
tpbTitle: Debian -- Details of package libargtable2-0 in jessie (at
tumbleweedthat patch looks like it parses ints unsigned, instead of signed?04:34
tumbleweedthat doesn't look very compatible with upstream04:34
tumbleweedwhy is it required?04:34
CarlFK "A makestuff build of Stewart Heitmann's argtable2, including an unsigned int patch which Stewart didn't wish to include."04:34
tpbTitle: makestuff/libargtable2 · GitHub (at
tumbleweedright, but why are we dependant on that patch?04:34
CarlFKdon't know, other than it is currently being built as part of the "make all the things" script to bring up the Atlys04:35
tumbleweedit's an API-incompatible patch - presumably if it had been more sensible, upstream would have accepted it04:35
CarlFKhere is what I have worked out that builds flcli
tpbTitle: Ubuntu Pastebin (at
tumbleweedbut basically, if you're dependant on a minor fork of a library, like that04:38
tumbleweedgetting a package into Debian is very very hard04:38
tumbleweedit's best to resolve the cause of the fork04:39
CarlFKnot surprised04:39
CarlFKwell.. for pyconza I would like to get this shoved into a ppa04:39
CarlFKin a somewhat done right way.04:40
CarlFKI am ok relying on apt-get from launchpad.   not so ok relying on wget sourceforge04:41
CarlFKmake/wget/patch srcript of doom...04:42
CarlFKright now, my build script (builds flcli  runs make.. which does rm -rf argtable2-${AT2_VERSION};  tar zxf argtable2.tgz04:44
tpbTitle: Ubuntu Pastebin (at
CarlFK I guess I can just # those 2 lines out, untar it, and then all the code is "there"04:44
CarlFKtar up the code/Makefiles and roll it into one big ugly package.04:45
tumbleweedso, you have 2 possible approaches here04:45
tumbleweedtar up everything into one big package, yeah04:45
tumbleweedor package each thing separately04:46
CarlFKpretty sure separately will require re-writing the Makefile(s)04:47
CarlFKtumbleweed: should ./configure be run as part of the "get source in order, then tar" or part of the build process?06:10
tpbTitle: flcli build script · Issue #74 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
CarlFKI am done for tonight.   Zzzzzz07:10
*** rohitksingh has joined #timvideos08:58
ysionneau03:25 <@mithro> CarlFK: I haven't figure out why - but it seems sometimes the build takes 2.5 hours - most of the time it takes 30 minutes... < same thing for ARTIQ project when building with ISE09:16
ysionneauwe don't have the issue with Vivado though09:17
*** rohitksingh has quit IRC11:17
*** f15h has joined #timvideos12:47
*** f15h has quit IRC13:03
tumbleweedCarlFK: build process13:59
CarlFKtumbleweed: k. thanks.14:00
*** rohitksingh has joined #timvideos15:28
*** rohitksingh has quit IRC17:54
*** rohitksingh has joined #timvideos18:09
*** Sewar has quit IRC20:56
*** Sewar has joined #timvideos20:57
*** rohitksingh has quit IRC21:19
CarlFKtumbleweed: do you have time in the next 24 hours to take a shot at packaging flcli (now that I have smoothed out all the wget/tar nonsense)21:24
CarlFKor do you want me to spend a few hours and then you can fix it21:24
CarlFK(ps, I don't really have a few hours)21:25
springermac_mithro: If you have a second do you remember what part of gst-python those patchs were to? or
*** CarlFK has quit IRC23:00

Generated by 2.13.1 by Marius Gedminas - find it at!