New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ipython console freezes for increasing no of seconds in %pylab mode #6189
Comments
Forgot to say, I'm driving Qt4Agg usually: n [1]: %pylab
Using matplotlib backend: Qt4Agg
Populating the interactive namespace from numpy and matplotlib Interestingly, going back to rel-2.1.0 does not solve this. |
I went back as far as 1.2 and it still happens in the terminal ipython. |
We've had some issues with the Qt backend readline interacting before, though I can't find the issue number just at the moment. |
This is a severe problem, and I have reproduced it with nothing more than
so it is not matplotlib-specific. To see it on Mavericks, with either Homebrew or Anaconda python 3.4.1 and PyQt4, just start a console as above, then go away and do something elsewhere for a few minutes. When you come back, chances are the session will be unusable because of delayed response to attempts to type or scroll through history. It's somewhat erratic, but qualitatively repeatable. In my case, testing is with a plain Apple terminal window (with focus follows mouse), not with ITerm2. |
@takluyver, The earlier issue involving qt and readline is #481. Despite the issue title, it involved more than just XP. More than one distinct problem was discussed in that issue, but the primary one involved the input hook for qt. |
A workaround is to use the qtconsole; at least in a little bit of testing, it seems to maintain full responsiveness. |
…n#6189 On Mavericks, the qt and wx gui support was unusable from the ipython console in a terminal. Using nope() in enable_qt4 and enable_wx, and reversing with nap() in disable_qt4 and disable_wx, solves the problem.
Has anyone seen this problem on Linux? @ellisonbg, @minrk, @jdfreder: you guys all have macs, can one of you shed any extra light on this? It's been a while since we've heard of issues with our event loop integration, and I can't think of any changes in that code recently, but perhaps something subtle happened... This sounds pretty serious. |
This is a long shot on being related, but matplotlib/matplotlib#3360. |
Not related; the problem appears without matplotlib. |
Thanks for tracking it down, I had forgotten about App Nap, and the fix we made in kernels, but not in terminal IPython. |
Use appnope in qt and wx gui support from the terminal; closes #6189
…n#6189 On Mavericks, the qt and wx gui support was unusable from the ipython console in a terminal. Using nope() in enable_qt4 and enable_wx, and reversing with nap() in disable_qt4 and disable_wx, solves the problem.
Great, glad this is also fixed in the terminal. Thanks!! |
…n#6189 On Mavericks, the qt and wx gui support was unusable from the ipython console in a terminal. Using nope() in enable_qt4 and enable_wx, and reversing with nap() in disable_qt4 and disable_wx, solves the problem.
launching ipython just in console with
ipython
.Am trying to reproduce bad editing (readline?) behavior now, looking at some functions, their help text etc. but it's all ok.
Then launching
%pylab
.Typing in rubbish (like 'aoeuaoeaoeu') and can keep editing it without freezes.
But from the moment I type in a function name that exists, like
dir()
it even gets stuck before printing the()
. From now on the cursor gets stuck every 5 seconds or so and keeps stuck for 2-3 seconds at a time, but that number is increasing.Version abffd96
inside iTerm2 on Mac OSX.
Python 2.7.8 from conda
The text was updated successfully, but these errors were encountered: