These are chat archives for ipython/ipython

15th
Oct 2014
Thomas Kluyver
@takluyver
Oct 15 2014 00:05
@bollwyvl bug in Chrome. It apparently works in 38 and 40
Nicholas Bollweg
@bollwyvl
Oct 15 2014 00:06
cool, good to know
Min RK
@minrk
Oct 15 2014 00:10
@bollwyvl patch has been approved for Chrome 39, should be fixed on the next Chrome beta update.
Nicholas Bollweg
@bollwyvl
Oct 15 2014 00:12
oh, flexbox... so much promise
Jonathan Frederic
@jdfreder
Oct 15 2014 00:18
@SylvainCorlay @jasongrout looks like @bollwyvl has the rendering issue I had when I came to visit.
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 00:30
yes. I am still under Chromium 0.37. Oddly Jason got that update as I did not and has the same rendering issue.
Dhruv Madeka
@dmadeka
Oct 15 2014 00:50
@takluyver Hello! I saw that you wrote the R Kernel, and noticed it didn't handle widgets. Quick question, did you think of using ActiveBindings to mimic traitlet behaviour in the R Kernel?
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 02:26
@jdfreder is there any reason to limit the version of lessc to 1.7.5 ?
1.7.6 is convenient as it comes with the last stable ubuntu
the only difference in what it generates is the "old browsers" section
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 02:42
(now absent)
Min RK
@minrk
Oct 15 2014 02:54
@SylvainCorlay 1.7.5 or 1.7.6 makes significant changes to output, causing a bunch of churn. We aren't sure if it was intentional or not, but we are checking on it.
How does 1.7.6 come with stable ubuntu? 1.7.5 is the most recent release
Min RK
@minrk
Oct 15 2014 03:17
ah I've narrowed it down. It's a bug in comment parsing. Workaround is easy, I'll investigate a proper fix, as well.
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 04:33
@minrk I meant 1.7.5. The version interval is exclusive.
on the right-hand side.
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 04:51
Arg I am sorry if you spent time on this. I just looked at the change in the rc file I did yesterday. As I said the upper bound if the interval is exclusive.
epifanio
@epifanio
Oct 15 2014 06:19
latest master fails to load notebooks with this error https://gist.github.com/d8bf96b86f339643b732
i’ll try a : python setup.py submodule
the error persist
epifanio
@epifanio
Oct 15 2014 06:32
refresh the page and a server restart, problem fixed .. i guess something changed in the nb format ?
now works
Matthias Bussonnier
@Carreau
Oct 15 2014 10:26
@epifanio no, server restart is required when submodule updates.
Damian Avila
@damianavila
Oct 15 2014 14:06
@rgbkrk, how temporary is tmpnb?
Kyle Kelley
@rgbkrk
Oct 15 2014 14:06
the notebook servers stay around for an hour (minus the ops shenanigans I've had to pull this morning)
There's a gitter room for tmpnb too btw
Damian Avila
@damianavila
Oct 15 2014 14:07
thanks...
Damian Avila
@damianavila
Oct 15 2014 15:57
RISE powered by tmpnb: https://tmpnb.org/user-A3k8XrwHF9Is/notebooks/RISE.ipynb
I know there is a tmpnb room but it is so nice to have this facility, so sharing the link here too... thanks again @rgbkrk
Kyle Kelley
@rgbkrk
Oct 15 2014 15:59
We'll probably be setting cookies for each notebook and letting you share as gist
Damian Avila
@damianavila
Oct 15 2014 16:00
yes... would be nice...
Matthias Bussonnier
@Carreau
Oct 15 2014 16:02
Ohhhhhh
Jonathan Frederic
@jdfreder
Oct 15 2014 16:04
:police_car: danger zone
Matthias Bussonnier
@Carreau
Oct 15 2014 16:05
ah sorry.
Jonathan Frederic
@jdfreder
Oct 15 2014 16:06
lol np, I was jk
Matthias Bussonnier
@Carreau
Oct 15 2014 16:08
:-P
Matthias Bussonnier
@Carreau
Oct 15 2014 17:27
@takluyver Can't seem to find reliable doc on AltGr vs Alt
And on Mac there is not AltGr, you do have to press Alt to get these characters.
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 18:01
hello,
We have some widgets that should not be directly displayed in another context than as a child of a certain widget.
Hence, calling display on them does not make sense.
I overload _ipython_display_(self, **kwargs) to just do print(str(self))in this case.
Is there a more elegant solution?
Jonathan Frederic
@jdfreder
Oct 15 2014 18:11
@SylvainCorlay I think overloading _ipython_display is the right thing to do
but maybe instead of print(str(self))
maybe call self.__repr__()
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 18:13
print(self.__repr__()) ?
Jonathan Frederic
@jdfreder
Oct 15 2014 18:13
looking*... I think there's even something better you can do
yeah
but I think I have a better idea
one sec
try _ipython_display_ = None in the class definition
I think that should let the display hook work like normal
I think that's cleaner because the IPython rich display framework would work.
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 18:17
_ipython_display = None does not trigger the printing of the __repr__ though
see here :)
@rgbkrk tmpnb is wonderful
:D
@SylvainCorlay do you see what I mean?
@SylvainCorlay It causes the custom logic to be bypassed and allows the display hook logic to run. See this line: https://github.com/ipython/ipython/blob/bfabda65d1434a1299ed789301dfe9bf7e7c652c/IPython/core/displayhook.py#L232
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 18:22
Yeah, it is perfect!
Jonathan Frederic
@jdfreder
Oct 15 2014 18:22
Glad I could help :)
Sylvain Corlay
@SylvainCorlay
Oct 15 2014 18:23
thanks, it is really what I was looking for
Remi Rampin
@remram44
Oct 15 2014 21:04
Alright, filed a super-weird error as #6709 that is super-annoying to me for super-bad reasons :(
Jonathan Frederic
@jdfreder
Oct 15 2014 23:43
Does this look familiar to anyone?
[I 16:41:46.523 NotebookApp] Kernel started: 6b20c16f-e602-4384-b2d5-b26e3e35b0db
Page Error!
line 2 of evaluate
line 3 of evaluate
line 3 of evaluate
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:43
Yeah
It has to do with something not being defined when the test is being run
Or, is this in a test even?
Jonathan Frederic
@jdfreder
Oct 15 2014 23:44
it is
thanks Jess
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:44
Ok, yeah
Jonathan Frederic
@jdfreder
Oct 15 2014 23:44
hmm
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:44
Yeah, the error message is annoyingly vague
You might be able to get some more hints if you add casper.print_log() to the end of util.js
Jonathan Frederic
@jdfreder
Oct 15 2014 23:45
TypeError: 'null' is not an object (evaluating 'IPython.notebook.kernel.is_connected')
not sure why the kernel isn't defined here
I think something isn't importing correctly
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:45
Hmm, yeah
Jonathan Frederic
@jdfreder
Oct 15 2014 23:46
Captured console.log:
    Default extension for cell metadata editing loaded.
    Raw Cell Format toolbar preset loaded.
    Slideshow extension for metadata editing loaded.
    Cell tags extension loaded.
    no checkpoint done
    Session: kernel_created (fb16c2c9-6443-4e92-9797-5a4a7700f25c)
    Starting WebSockets: ws://127.0.0.1:8888/api/kernels/6b20c16f-e602-4384-b2d5-b26e3e35b0db
    Unexpected response code: 426
    Unexpected response code: 426
    Unexpected response code: 426
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:47
Ah, now that’s a phantom thing I think
At least, I encountered it when I was trying to test a PATCH request
Which phantom doesn’t support yet apparently
Jonathan Frederic
@jdfreder
Oct 15 2014 23:48
hmm... This is the bootstrap-tags PR
I don't actually add any handlers
Just a new component
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:48
Hmm
Jonathan Frederic
@jdfreder
Oct 15 2014 23:48
but I don't remember having trouble doing that in the past
I think I'm going to solve this problem by pulling in parts one by one in a new branch
see what causes it to break
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:50
Yeah, probably a good idea
I actually have to run or I’d try to help
But good luck figuring it out!
Jonathan Frederic
@jdfreder
Oct 15 2014 23:51
Thanks
!
I'll let you know what it is
Jessica B. Hamrick
@jhamrick
Oct 15 2014 23:51
Cool, thanks!