sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
Gurty has joined #m-labs
Gurty has joined #m-labs
_whitelogger has joined #m-labs
fengling has joined #m-labs
sb0 has joined #m-labs
<cr1901_modern> sb0: Oops... looks like I asked you a q while you weren't in the room lmao. Are you around?
nicksydney has quit [Remote host closed the connection]
<ysionneau> cr1901_modern: he usually looks at irc logs on the web (cf topic)
fengling has quit [Quit: WeeChat 1.1.1]
<GitHub36> [artiq] enjoy-digital pushed 3 new commits to master: http://git.io/vL0pX
<GitHub36> artiq/master 38a0f63 Florent Kermarrec: gateware/soc: use Minicon SDRAM controller and 128KB shared L2 cache
<GitHub36> artiq/master 449964c Florent Kermarrec: runtime/mailbox: remove flush of L2 cache (L2 cache is now shared between CPUs)
<GitHub36> artiq/master d25a07f Florent Kermarrec: Merge branch 'master' of https://github.com/m-labs/artiq
<cr1901_modern> ysionneau: Oops...
<sb0> cr1901_modern, sure, go ahead and send it
<sb0> but i don't use windows, sending it to the list enables windows users (e.g. Florent) to have a look at it
<GitHub81> [artiq] fallen pushed 1 new commit to master: http://git.io/vLurk
<GitHub81> artiq/master 77ca8bb Yann Sionneau: artiq_coreconfig: better arg parsing
sb0 has quit [Quit: Leaving]
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#224 (master - 77ca8bb : Yann Sionneau): The build has errored.
travis-ci has left #m-labs [#m-labs]
<GitHub41> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vLzTW
<GitHub41> artiq/master 3636025 Sébastien Bourdeauducq: pipistrello: smaller L2 cache
<ysionneau> sb0 : there "data" is both a method and the backing store, OK to rename the latter as "backing_store"? (I'm working on the search field for parameters, I have something working but I had to rename the backing store so that I could call the data() method)
<ysionneau> here is what I have so far (renaming of data to backing_store + working search field): https://github.com/fallen/artiq/commit/e4c2953f41fdbf6a8762ced5f303eab1e85bc18c
<ysionneau> the Model/View of Qt seems a bit like overengineering, when you first read all the code necessary to just display a Table
<ysionneau> but it feels clean at least
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#225 (master - 3636025 : Sébastien Bourdeauducq): The build passed.
travis-ci has left #m-labs [#m-labs]
travis-ci has joined #m-labs
<travis-ci> fallen/artiq#239 (gui - bc44a9e : Yann Sionneau): The build passed.
travis-ci has left #m-labs [#m-labs]
<GitHub70> [artiq] fallen pushed 1 new commit to master: http://git.io/vLg4X
<GitHub70> artiq/master 1ad5165 Yann Sionneau: artiq_gui: add a search box in parameters dock
<GitHub119> [artiq] fallen pushed 1 new commit to master: http://git.io/vLgBk
<GitHub119> artiq/master b059720 Yann Sionneau: artiq_gui: check due date upon date selection
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#226 (master - 1ad5165 : Yann Sionneau): The build passed.
travis-ci has left #m-labs [#m-labs]
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#227 (master - b059720 : Yann Sionneau): The build passed.
travis-ci has left #m-labs [#m-labs]
sb0 has joined #m-labs
<sb0> ysionneau: coreconfig write should print an error if no -s/-f is specified
<ysionneau> ack
<sb0> is there no way of making the left-click menu simpler, e.g. connect a signal?
<ysionneau> we could try to emit the ContextMenuEvent upon left-click (.clicked signal)
<ysionneau> but label does not have any clicked signal
<ysionneau> but we can inherit from label and add support for left click and raise the ContextMenuEvent, but then it's not really simpler :/
<sb0> that's what you've done, right?
<ysionneau> ah right sorry
<sb0> $ lda_controller -d 368 Segmentation fault (core dumped)
<ysionneau> there is a linkActivated() signal on QLabel, I guess if we put a link in the text .... but that would be some kind of hack
<ysionneau> sb0: what happens if you don't specify "-d"?
<sb0> it runs
<sb0> it also crashes if i specify a serial number that does not exist
<ysionneau> I think we already talked about this, you need to specify SN:0368, and you also told me to put it in the doc, which I thought I did ... but did not :o
<sb0> lda_controller -d SN:0368 Segmentation fault (core dumped)
<sb0> and it should not crash, anyway
<ysionneau> yes
<ysionneau> I should also fix that crash
<ysionneau> SN:00368
<ysionneau> ?
<sb0> crash
<ysionneau> rha
<sb0> if self.serial is None: self.serial = next(self.enumerate(self.product))
<sb0> how do I use this feature of the driver when running the controller?
<ysionneau> this feature? This just takes the first device if you don't specify a serial number
<sb0> ysionneau: do i need specific udev rules for the permission?
<sb0> ysionneau: yes and "lda_controller: error: argument -d/--device: expected one argument"\
<sb0> ysionneau: hence my somewhat rhetorical question.
<sb0> ysionneau: and when calling enumerate manually, it returns None
<sb0> ysionneau: now in thorlabs_tcube_controller
<sb0> AttributeError: 'TdcSim' object has no attribute 'close'
<ysionneau> allright, the segfault is due to not having correct rights
<ysionneau> I have same issue, but with sudo it works
<ysionneau> with wrong device ID it raises IOError, with correct device ID it works
<sb0> ysionneau: is there a way the tcube stuff can use SI units instead of "encoder counts", or is that excessively complicated considering the variety of actuators that can be connected to it?
<ysionneau> I don't remember quite well the documentation, would need to check that
<sb0> ysionneau: also, when you have multiple tcube devices connected to a USB hub, how do you identify them by serial number?
<ysionneau> don't you still receive the usb serial number, even connected through a USB hub?
<ysionneau> it's in the descriptor AFAIK
<sb0> ysionneau: can you use udev to map that to a consistent entry in /dev instead of a somewhat random ttyUSBX?
<ysionneau> ah sorry I was still on the lda topic
<ysionneau> yes udev can do that
<sb0> ysionneau: for the tcube? can you add the procedure to the documentation?
<ysionneau> to tie some usb uart device to always the same ttyUSBX
<ysionneau> I haven't done it yet, but I will have a look and add it to the doc when I have a working udev rule file
* ysionneau still does not understand why lda controller segfaults when not run in root
<ysionneau> why doesn't it just print some access denied ...
<ysionneau> looks like hid_open is causing the segfault
<ysionneau> seems like a libhidapi-libusb bug ... I compiled their example on their website http://www.signal11.us/oss/hidapi/
<ysionneau> and it produces a segfault as well when not run as root
<ysionneau> ah but they don't do any check on their example
<ysionneau> ok let me have a closer look
<ysionneau> maybe the ctypes binding is wrong...
<ysionneau> a simple C example which does hid_open() as a non-root user does not segfault, it returns NULL as stated in the doc.
<ysionneau> so the issue must be in the ctypes or the calling (maybe the 3rd argument, which is a wide char string)
<ysionneau> which seems to be the case (3rd argument issue) since segfault happens at wcscmp () at ../sysdeps/x86_64/wcscmp.S:426
<sb0> ysionneau: how can I wait until a tcube move_relative call is done moving the motor?
<sb0> ysionneau: i see that it is implemented as "async" in move_jog. why don't other functions have this option?
<sb0> ysionneau: also, if possible, do not implement the non-waiting variant. there is the general fire_and_forget option in artiq for dealing with that.
<cr1901_modern> sb0: Good point. I did in fact post it to the mailing list for others to check. I keep forgetting florent uses Windows as well. Just don't want to bog down the mailing list with trivial changes.
<ysionneau> sb0: hum, I thought it was blocking until the end of the move, but it's not, I should fix that
<ysionneau> (move_relative)
<sb0> all methods. and remove that async parameter from jog.
<ysionneau> I can wait for the ENDOFMOVEMSG
<ysionneau> ok
<sb0> whatever is clean and reliable.
<ysionneau> ack
rjo has quit [Quit: leaving]
<ysionneau> in fact I can just wait for the MOT_MOVE_COMPLETED
<sb0> ysionneau: the correct string for lda is SN:00368
<sb0> this should be documented.
<sb0> I suggest adding it to -h as well
<ysionneau> ok, this time I won't forget to do it.
rjo has joined #m-labs
<GitHub93> [migen] enjoy-digital pushed 3 new commits to master: http://git.io/vL2Qq
<GitHub93> migen/master 6370acd William D. Jones: Xilinx Platforms now use cmd.exe on Windows instead of bash to run scripts...
<GitHub93> migen/master 743a5f6 Florent Kermarrec: mibuild/xilinx/ise: simplify default_ise_path
<GitHub93> migen/master 7d8f4d1 Florent Kermarrec: mibuild/xilinx/ise: fix source and set source to False by default on Windows (tools supposed to be in the PATH)
<sb0> ysionneau: can you document how to use hwgrep to open a tcube device by serial number? and check that it actually works ...
<sb0> if that works, then we don't need udev
travis-ci has joined #m-labs
<travis-ci> m-labs/migen#42 (master - 7d8f4d1 : Florent Kermarrec): The build passed.
travis-ci has left #m-labs [#m-labs]
<sb0> ysionneau:
<sb0> Traceback (most recent call last): File "/home/rabi2/anaconda3/lib/python3.4/site-packages/artiq/gui/explorer.py", line 104, in submit_clicked expinfo = self.explist_model.data[key] TypeError: 'method' object is not subscriptable
ysionneau has quit [*.net *.split]
<GitHub191> [artiq] fallen pushed 1 new commit to master: http://git.io/vL2Fy
<GitHub191> artiq/master 4d077f5 Yann Sionneau: thorlabs tdc driver: all moves are now blocking and wait for move to end or be stopped
ysionneau has joined #m-labs
<GitHub140> [artiq] fallen pushed 1 new commit to master: http://git.io/vL2NK
<GitHub140> artiq/master 645c650 Yann Sionneau: gui: forgot to rename one data to backing_store
<ysionneau> need to go, will work on all the reported issues tomorrow morning
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#229 (master - 645c650 : Yann Sionneau): The build passed.
travis-ci has left #m-labs [#m-labs]