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
balrog has quit [*.net *.split]
balrog has joined #m-labs
bentley` has quit [Remote host closed the connection]
bentley` has joined #m-labs
<sb0> the new scan widget looks great!
<sb0> so what is delaying 1.0rc1 now: compiler/runtime issues ( whitequark ), and QC2 confirmation/docs from Daniel
<sb0> rjo, have you looked much into why the DDS is so slow?
<GitHub101> [artiq] sbourdeauducq pushed 1 new commit to master: https://git.io/vaapR
<GitHub101> artiq/master 633bbc4 Sebastien Bourdeauducq: gui/schedule: work around Qt bug when first row is inserted with due date. Closes #329
<sb0> whitequark, can you get those pesky Qt bugs properly fixed? with Aleksey's help maybe...
<sb0> also, the log resizing is still slow
<sb0> and the layout breaks when the horizontal size is small.
<sb0> Qt's theoretical behavior of displaying a horizontal scroll bar instead of wrapping long words is fine by me
<sb0> the slowness is probably qt/pyqt's fault. afaict all it takes is to set a model that has more than a few rows...
<sb0> even if that model returns empty data
<sb0> yes, it is. and it's *layout* options that cause the slowness
<sb0> let me bisect which one exactly...
<sb0> it's horizontalHeader().setSectionResizeMode(QtWidgets.QHeaderView.ResizeToContents)
<sb0> not only is automatic qt table layout lousy with bugs, it's also slow
fengling has quit [Ping timeout: 240 seconds]
fengling has joined #m-labs
<sb0> hm, it's not the only factor
FabM has joined #m-labs
<sb0> https://github.com/z3ntu/QtWaitingSpinner sometimes segfaults on shutdown, I guess due to the shobby pyqt memory management with timers ...
<sb0> *shoddy
<rjo> sb0: yes. i would not let the "waiting" issues delay rc1.
<rjo> sb0: all i noticed that the dds setting time can even be as long as 200 us (on pipistrello) and it seems to fluctuate a lot.
<rjo> sb0: i was unable to determine wether it is actually wrong timestamps or just too much cpu time being lost.
<sb0> the analyzer should give the timestamps
<sb0> rjo, whitequark, shall I take care of the remote imports on controllers?
<sb0> and an opencv/video controller sounds like a good demo
<rjo> sb0: remote imports and the datacrunching thing?
<sb0> yes
<rjo> sb0: if you end up wrapping a camera api, i would try to stay away from opencv.
<rjo> sb0: i had written parts of pydc1394 and pyflycapture2 (windows people love it and the cameras)
<rjo> sb0: and i have a API wrapper for the andor cameras.
<sb0> can i easily demonstrate the remote import of a function that does e.g. face recognition?
<rjo> that may look cool but is not useful in the lab ;)
<sb0> yes
<sb0> what would be useful in the lab, and semi-trivial to implement?
<rjo> the neutrals would be convinced by the demonstration that you can fit a 2D gaussian to noise 2D gaussian data.
<rjo> and they would be really impressed if you can fit a fermi profile (with a nasty and expensive dilogarithm) to noise fermi profile data.
<sb0> is there copy-paste/library code for that?
<rjo> the former has been done many times. there must be github code "bec fit gaussian parabola thomas fermi thermal" would be a set of search words.
<rjo> for dilog/polylog it is a bit more involved but would be nice because it would showcase linking c libraries (in this case gsl) with python and the data crunching.
<rjo> and potentially even dynamic compilation (numba/cython/weave...)
<rjo> i wrote code for that a couple years ago (weave+gsl).
stekern has quit [Ping timeout: 244 seconds]
stekern has joined #m-labs
<sb0> whitequark, how do I get a QScrollArea to expand its child widget?
<sb0> child.setSizePolicy(QtWidgets.QSizePolicy.Expanding, QtWidgets.QSizePolicy.Expanding) = nix
<sb0> child.expandingDirections returning Qt.Horizontal | Qt.Vertical = nix
<sb0> both = nix
<sb0> Qt layout sytem = steaming crap
<sb0> ah
<sb0> scroll_area.setWidgetResizable(True)
<sb0> phew
<GitHub168> [artiq] sbourdeauducq pushed 1 new commit to master: https://git.io/vaVuW
<GitHub168> artiq/master f739341 Sebastien Bourdeauducq: gui/moninj: flow layout
<rjo> sb0: that just created a single column of huge buttons for me.
<rjo> well not always single column but huge ttl "buttons"
<rjo> but it does flow.
<sb0> so, basically they're too large?
<sb0> can you send a screenshot?
<rjo> yes. 350x300 is too large.
<sb0> also they shouldn't look so much like buttons now?
<rjo> no. they look like "fields" or "areas". that's fine. http://imgur.com/CmrMohq
<rjo> i suspect their size should be relative to the text size (avgCharWidth() and lineSpacing())
<sb0> rjo, better now?
<GitHub33> [artiq] sbourdeauducq pushed 1 new commit to master: https://git.io/vaVaf
<GitHub33> artiq/master fb8a548 Sebastien Bourdeauducq: gui/moninj: compute size of widgets based on font sizes
<rjo> good. for me personally, the ttls should still be a factor 9 smaller (in area) and the dds a factor of 4.
<rjo> how do i make a QWidget show (visually, that blue tint) that it has focus?
rohitksingh has joined #m-labs
rohitksingh1 has joined #m-labs
rohitksingh has quit [Ping timeout: 240 seconds]
cr1901_modern has quit [Ping timeout: 244 seconds]
<GitHub87> [artiq] sbourdeauducq pushed 2 new commits to master: https://git.io/vaw11
<GitHub87> artiq/master 4bf0db3 Sebastien Bourdeauducq: master: publish repository status
<GitHub87> artiq/master 2859382 Sebastien Bourdeauducq: gui: display repository scanning status and revision. Closes #274
cr1901_modern has joined #m-labs
rohitksingh has joined #m-labs
rohitksingh1 has quit [Ping timeout: 246 seconds]
FabM has quit [Quit: ChatZilla 0.9.92 [Iceweasel 38.7.0/20160308234001]]
rohitksingh has quit [Ping timeout: 240 seconds]
_whitelogger has joined #m-labs
bentley` has quit [Ping timeout: 240 seconds]
fengling has quit [Ping timeout: 240 seconds]
bentley` has joined #m-labs
fengling has joined #m-labs
fengling has quit [Ping timeout: 240 seconds]
<GitHub131> [artiq] jordens pushed 1 new commit to master: https://git.io/var7J
<GitHub131> artiq/master 33da27a Robert Jordens: experiment/gui: support reverse scan (closes #246)
<GitHub47> [artiq] jordens pushed 1 new commit to master: https://git.io/vaoUz
<GitHub47> artiq/master f816c7d Robert Jordens: scanwidget: add shift-lmb and ctrl-lmb drag modes (1d78c78)
Gurty has quit [Ping timeout: 276 seconds]