00:05 <travis-ci> m-labs/artiq#72 (master - f4cb97b : Robert Jordens): The build passed.

00:05 <travis-ci> Build details : http://travis-ci.org/m-labs/artiq/builds/55883133

20:05 <rjo> ysionneau: if you are fighting with your travis-ci build, I would recommend debugging it in a chroot (they are quickly created). or you might want to have a look at circle-ci which would even give you ssh access into a build.

20:05 <rjo> ysionneau: but otherwise my impression is that most of your branches are ready to be merged. the small issues can be sorted out later.

20:44 <ysionneau> rjo: it's almost OK indeed :)

20:47 <ysionneau> only issue left in head is: on a freshly installed machine, if someone wants to try out ARTIQ using conda (which is the main goal), doing artiq_flash.sh will fail because you don't have access to the USB device

20:47 <ysionneau> and you most likely haven't set up the udev rule

20:48 <ysionneau> and just sudo artiq_flash.sh does not work because of miniconda env ... what you need to do is sudo -E PATH=$PATH artiq_flash.sh which is not very pretty to put in the web page as "simple install instructions" :/

21:24 <GitHub163> [misoc] enjoy-digital pushed 2 new commits to master: http://git.io/jnK3

21:24 <GitHub163> misoc/master e79a716 Florent Kermarrec: software/bios/sdram: select the type of data we want to generate for memtest with TEST_RANDOM_DATA (debugging hardware is easier with a simple counter)

21:24 <GitHub163> misoc/master e6de4b1 Florent Kermarrec: sdram/phy: add simphy (software memtest OK in simulation with MT48LC4M16)

22:13 <rjo> ysionneau: shipping the udev rule as a file and just telling them to sudo cp misc/artiq.rules /etc/udev/rules.d/70_artiq.rules is the easiest imho.

22:13 <ysionneau> ah, yes that's a solution

22:13 <rjo> ysionneau: especially if you make this hint a message of flash.sh if xc3sprog failes.

22:14 <rjo> fails.

22:18 <ysionneau> good, it seems xc3sprog returns 1 if "permission denied" and 3 for flashing errors :)

22:26 <rjo> and we should sprinkle some xc3sprog -R around

22:26 <ysionneau> ah yes at the end of the script maybe

22:28 <GitHub15> [misoc] enjoy-digital pushed 3 new commits to master: http://git.io/jceV

22:28 <GitHub15> misoc/master 9a9af17 Florent Kermarrec: sdram/phy/simphy: remove use of iter

22:28 <GitHub15> misoc/master 38d24b6 Florent Kermarrec: software/bios/sdram: make seed_to_data static

22:28 <GitHub15> misoc/master 9137b91 Florent Kermarrec: sdram: remove nbits from modules and databits from GeomSettings

22:39 <GitHub91> [misoc] enjoy-digital pushed 1 new commit to master: http://git.io/jcUJ

22:39 <GitHub91> misoc/master 340014d Florent Kermarrec: targets: revert use of integers in clocks/timings

23:17 <ysionneau> rjo: since we often change migen stuff, I think we cannot merge the artiq conda stuff yet, before having also automated generation of migen conda package

23:17 <ysionneau> so that latest artiq builds with latest migen, else it might fail to build when changing migen api

23:19 <sb0> we shouldn't break the migen api very often...

23:20 <ysionneau> ok, then we might be ok with merging then, latest commit is running in travis at this moment

23:37 <rjo> i would either fix the migen/misoc dependency of artiq and use conda packages for them or not use conda packages for migen/misoc and get from from git as i had it before.

23:38 <rjo> but do merge it. then florent also gets free testing of misoc implicitly.

23:40 <rjo> and it might be interesting to start building not only for ppro but also pipistrello. that tests the ddram parts of misoc and also tests a larger artiq (16 ttl etc).

23:40 <rjo> on travis we can run the (unittests, runtime+bios+ppro, runtime+bios+pipistrello) triple in parallel

23:41 <rjo> sb0: if the migen api is stable, you should tag and make a release! ;)

23:44 <rjo> (with "fix" i meant "freeze").

23:55 <rjo> ysionneau: nice. don't you also have do do udevadm control --reload-rules?