These are chat archives for ipython/ipython

11th
Dec 2014
Candide Kemmler
@ckemmler
Dec 11 2014 12:11
Hi, I just installed iPython using apt-get install on trusty and --version gives me 1.2.1 -- when stable seems to be 2.3. pip install --upgrade ipython fails miserably
should I start again from scratch using the documented method
?
Matthias Bussonnier
@Carreau
Dec 11 2014 15:52
On which distrib are you ?
@ckemmler you can pip-install --user to install only for your user not system whide
Jason Grout
@jasongrout
Dec 11 2014 16:39
@jdfreder, @ellisonbg: the warning when you import widgets I think is too much
especially if we are splitting off widgets to a separate repo and project after IPython 3.0
The warning says, "WARNING: The widget API is still considered experimental and
may change by the next major release of IPython." I think it's a given that there can be backwards incompatible changes in major releases. Especially if the widget package is split off from IPython
Min RK
@minrk
Dec 11 2014 17:16
I'll host from here today
here being home, which I realize I should also mention, since none of you are also here
Matthias Bussonnier
@Carreau
Dec 11 2014 17:18
Then I'm here to, but unlike you as usual.
Jessica B. Hamrick
@jhamrick
Dec 11 2014 17:24
Hah, I guess I’m here too then :)
Thomas Kluyver
@takluyver
Dec 11 2014 17:24
crazy. I'm here, and I can't see any of you. Are you all under the bed?
Jessica B. Hamrick
@jhamrick
Dec 11 2014 17:25
I guess nobody (in Berkeley) wants to go out in the storm?
Thomas Kluyver
@takluyver
Dec 11 2014 17:26
so far, the "storm" just looks like ordinary rain from here. But the TV is busy saying that it's life threatening, and I should on no account venture outside, so...
Min RK
@minrk
Dec 11 2014 17:27
I heard some labs on campus have lost power
Jessica B. Hamrick
@jhamrick
Dec 11 2014 17:27
Haha, I haven’t even been looking at the tv
Oh, huh
Min RK
@minrk
Dec 11 2014 17:27
But yeah, right now it looks like 'pretty wet'
maybe there was a lot of wind over night or something
Brian E. Granger
@ellisonbg
Dec 11 2014 17:28
Oh, given the news, I thought Berkeley had been consumed by apocolyptic rain, protests, etc and vanished altogether
Thomas Kluyver
@takluyver
Dec 11 2014 17:29
maybe the New World Order is controlling the weather to damp down protests
Min RK
@minrk
Dec 11 2014 17:29
probably
Brian E. Granger
@ellisonbg
Dec 11 2014 17:29
LOL - yes
Jessica B. Hamrick
@jhamrick
Dec 11 2014 17:29
Hahaha
Thomas Kluyver
@takluyver
Dec 11 2014 17:29
in the UK, it's widely rumoured that unpopular legislation is pushed in the winter because people are less willing to protest.
Brian E. Granger
@ellisonbg
Dec 11 2014 17:30
Although, I think the New World Order would figure out how to just make it rain on black people
Min RK
@minrk
Dec 11 2014 17:30
"Californians fear the rain, we shall dampen them and thereby their spirits!"
Brian E. Granger
@ellisonbg
Dec 11 2014 17:30
But I can imagine the various police departments are quite relieved
Thomas Kluyver
@takluyver
Dec 11 2014 17:31
yeah
Min RK
@minrk
Dec 11 2014 17:31
Assuming they don't have to go stand out in the rain anyway.
Thomas Kluyver
@takluyver
Dec 11 2014 17:31
the other day there were two helicopters and one plane circling
Brian E. Granger
@ellisonbg
Dec 11 2014 17:31
@fperez said he was going to stay home, but was going to join in
Jonathan Frederic
@jdfreder
Dec 11 2014 17:31
Hey guys, running through the rain, will coonnect in 5
Fernando Perez
@fperez
Dec 11 2014 17:31
yup
just opened up...
Damian Avila
@damianavila
Dec 11 2014 17:32
I am not getting the invitation, can somebody post the link?
Fernando Perez
@fperez
Dec 11 2014 17:32
it's raining pretty hard here, but no serious damage.
but I was worried that the commute back in the evening would be hell/impossible. LBL sent out a warning to all employees to stay home.
Brian E. Granger
@ellisonbg
Dec 11 2014 17:32
Still pretty mild down here
Fernando Perez
@fperez
Dec 11 2014 17:32
I have to drive through hills/woods quite a bit, and if lots of trees go down, that could turn into hours stuck.
Min, can you run the meeting?
Brian E. Granger
@ellisonbg
Dec 11 2014 17:33
Yep - but your entire neighborhood will probably slide off the hill though
Damian Avila
@damianavila
Dec 11 2014 17:33
big storm?
Fernando Perez
@fperez
Dec 11 2014 17:33
Or someone on a mac and a good line?
yes @ellisonbg :)
yup @damianavila, huge storm in Northern California.
Brian E. Granger
@ellisonbg
Dec 11 2014 17:33
I though Min already started - but I can iff needed
Fernando Perez
@fperez
Dec 11 2014 17:33
ah!
is the link above?
Brian E. Granger
@ellisonbg
Dec 11 2014 17:33
Min did you start?
Matthias Bussonnier
@Carreau
Dec 11 2014 17:33
No
And with your rain are you sending me a message you don't want mein the US ?
Brian E. Granger
@ellisonbg
Dec 11 2014 17:35
I will say that in general CA is a bit narcissistic - our storms are definitely more important/better than anyone elses, just like our weather, startups, etc.
Damian Avila
@damianavila
Dec 11 2014 17:35
jajaaj
Jessica B. Hamrick
@jhamrick
Dec 11 2014 17:36
Yeah, given that on the east coast, a bad storm == hurricane, this doesn’t seem too bad in comparison :)
Scott Sanderson
@ssanderson
Dec 11 2014 17:36
There was flooding in boston yesterday and no one particularly noticed...
the T was slower than usual, but that's pretty much to be expected
Brian E. Granger
@ellisonbg
Dec 11 2014 17:37
LOL yeah - but here in CA, we could get earthquakes, storms, drought and protests at the same time :worried:
Fernando Perez
@fperez
Dec 11 2014 17:37
agreed :) so far this is basically what normal people describe as 'rain'
Jonathan Frederic
@jdfreder
Dec 11 2014 17:37
is the meeting going?
I don't see a link
Matthias Bussonnier
@Carreau
Dec 11 2014 17:37
@minrk forgot
Fernando Perez
@fperez
Dec 11 2014 17:37
ah, is Min around at all?
Matthias Bussonnier
@Carreau
Dec 11 2014 17:37
He is "here"
Jonathan Frederic
@jdfreder
Dec 11 2014 17:38
he chatted 6 min ago
Fernando Perez
@fperez
Dec 11 2014 17:38
I can try to start it, but I don't trust Chrome on linux to run a meeting stably.
Brian E. Granger
@ellisonbg
Dec 11 2014 17:38
he was
Thomas Kluyver
@takluyver
Dec 11 2014 17:38
yes, he said he'd run the meeting
Fernando Perez
@fperez
Dec 11 2014 17:38
ok
Brian E. Granger
@ellisonbg
Dec 11 2014 17:38
Let me try
Matthias Bussonnier
@Carreau
Dec 11 2014 17:38
Reviewing PR
Fernando Perez
@fperez
Dec 11 2014 17:38
tx
Matthias Bussonnier
@Carreau
Dec 11 2014 17:39
he is responding quicker on the PR than on github
let's open an issue "run meeting" and assign to @minrk
Min RK
@minrk
Dec 11 2014 17:39
Starting it up now
Damian Avila
@damianavila
Dec 11 2014 17:39
jajaja
Min RK
@minrk
Dec 11 2014 17:39
Sorry, got distracted
Matthias Bussonnier
@Carreau
Dec 11 2014 17:39
You see it worked !
Fernando Perez
@fperez
Dec 11 2014 17:41
link?
Brian E. Granger
@ellisonbg
Dec 11 2014 17:41
Where do I find the link again?
oh, I guess we have two of them...
Thomas Kluyver
@takluyver
Dec 11 2014 17:43
which one are we using
Matthias Bussonnier
@Carreau
Dec 11 2014 17:43
min's
Min RK
@minrk
Dec 11 2014 17:44
Sigh
Thomas Kluyver
@takluyver
Dec 11 2014 17:44
'This call is already full'
Scott Sanderson
@ssanderson
Dec 11 2014 17:45
@takluyver I can drop out if need be, I'm mostly just listening in
Thomas Kluyver
@takluyver
Dec 11 2014 17:45
are there actually 10 people? or is Google just being silly?
Damian Avila
@damianavila
Dec 11 2014 17:45
silly... we are 9
Matthias Bussonnier
@Carreau
Dec 11 2014 17:46
Kyle will jump out
Thomas Kluyver
@takluyver
Dec 11 2014 17:46
ok, let's try again
Kyle Kelley
@rgbkrk
Dec 11 2014 17:46
youtube link?
Damian Avila
@damianavila
Dec 11 2014 17:46
I can jump out too... if you have something to add @rgbkrk , just let me know...
Kyle Kelley
@rgbkrk
Dec 11 2014 17:47
sounds good
i like joining the event as an audience, I get +1 and -1 buttons
Thomas Kluyver
@takluyver
Dec 11 2014 17:49
for the whole thing, or do you get to rate us second-by-second?
Kyle Kelley
@rgbkrk
Dec 11 2014 17:51
yay
no idea
Scott Sanderson
@ssanderson
Dec 11 2014 18:04
what you just described is a lot like what we do for collaborative debugging on quantopian
which works well
Brian E. Granger
@ellisonbg
Dec 11 2014 18:04
ok
Kyle Kelley
@rgbkrk
Dec 11 2014 18:04
No one can be in the North East
Rather, noreasters can :bee:
Jason Grout
@jasongrout
Dec 11 2014 18:13
talking about warning boxes -- I'd saw we should limit the widget warnings to a big box in the documentation, but not a warning on import...
I'd suggest
Jonathan Frederic
@jdfreder
Dec 11 2014 18:14
That sounds like a good idea
I've already grown tired of the widget warning I added
Matthias Bussonnier
@Carreau
Dec 11 2014 18:14
noone read the doc.
Brian E. Granger
@ellisonbg
Dec 11 2014 18:16
I think it is important for all users to see this warning
Jason Grout
@jasongrout
Dec 11 2014 18:16
I think we should keep in mind doing something similar to what Thomas is doing with the prefixing stuff for nbextensions
Brian E. Granger
@ellisonbg
Dec 11 2014 18:16
Let's add to agenda though
Jonathan Frederic
@jdfreder
Dec 11 2014 18:18
I added @jasongrout 's comment to the agenda
Jason Grout
@jasongrout
Dec 11 2014 18:18
Considering that we would be splitting out widgets from base IPython after 3.0, they literally are disappearing before the next major release.
perhaps that's a reason to split them out before....
(it will be tough to backport fixes, etc., with the code being in two different repos)
so @jdfreder: you can change my agenda item to: splitting out widgets before 3.0. Unless you want to talk about it a bit before then. If needed, I can hop on using my phone...
Jonathan Frederic
@jdfreder
Dec 11 2014 18:27
I'll add both
Jason Grout
@jasongrout
Dec 11 2014 18:36
lots of systems use pkg-config to find various settings a package was installed by
(don't know if that is helpful in this prefix discussion, though)
the naming convention is like git, right?
git-mypackage is invoked when you do git mypackage, IIRC
Min RK
@minrk
Dec 11 2014 18:39
@jasongrout yes, exactly
Jason Grout
@jasongrout
Dec 11 2014 18:40
so how do you get a catalog of kernels?
Fernando Perez
@fperez
Dec 11 2014 18:40
jupyter-kernel-<TAB> :)
Jason Grout
@jasongrout
Dec 11 2014 18:40
oh great, using pexpect? :)
Brian E. Granger
@ellisonbg
Dec 11 2014 18:41
Programmaticaly, just use glob on all of the dirs on PATH
Jason Grout
@jasongrout
Dec 11 2014 18:41
huh, that makes it trivial to override system installs
Matthias Bussonnier
@Carreau
Dec 11 2014 18:41
That's just crazy, we should just think, sleep on it, and move on for now.
Jason Grout
@jasongrout
Dec 11 2014 18:42
+1 @Carreau
to thinking, sleeping, and moving on, anyway
Fernando Perez
@fperez
Dec 11 2014 18:42
kyle!
Kyle Kelley
@rgbkrk
Dec 11 2014 18:42
hi
Fernando Perez
@fperez
Dec 11 2014 18:42
can you hop in?
Kyle Kelley
@rgbkrk
Dec 11 2014 18:42
yeah
Jason Grout
@jasongrout
Dec 11 2014 18:46
domain idea: idontlikeipython.org
I tried to get on, but the call is full
Fernando Perez
@fperez
Dec 11 2014 18:47
:)
do you want to get on? we can ask someone who's not talking much to follow on youtube
Jason Grout
@jasongrout
Dec 11 2014 18:48
sure, I'd like to get on
Fernando Perez
@fperez
Dec 11 2014 18:48
ok
hold on
Jason Grout
@jasongrout
Dec 11 2014 18:48
present a case for splitting widgets off before 3.0
Fernando Perez
@fperez
Dec 11 2014 18:48
ok
Jessica B. Hamrick
@jhamrick
Dec 11 2014 18:49
I can drop out
Fernando Perez
@fperez
Dec 11 2014 18:49
sure?
up to you Jess
Damian and Scott have also been quiet, so I can ask them too...
ok, Jess did drop out. Jason, want to jump in?
thx jess!
Jessica B. Hamrick
@jhamrick
Dec 11 2014 18:49
No problem
Scott Sanderson
@ssanderson
Dec 11 2014 18:49
I'm happy to drop out as well if need be
Fernando Perez
@fperez
Dec 11 2014 18:49
Jess, did you talk to the math folks?
Matthias Bussonnier
@Carreau
Dec 11 2014 18:50
Ho, we already miss you @jhamrick !
Damian Avila
@damianavila
Dec 11 2014 18:50
yes, let me know... I can drop out if more places are needed...
Kyle Kelley
@rgbkrk
Dec 11 2014 18:51
@jasongrout did you already buy idontlikeipython.org
Fernando Perez
@fperez
Dec 11 2014 18:51
we need to find a solution to the max10 person problem
ideas?
Kyle Kelley
@rgbkrk
Dec 11 2014 18:51
Relocate people to San Antonio, TX
Fernando Perez
@fperez
Dec 11 2014 18:51
does appear.in support more than 10?
Kyle Kelley
@rgbkrk
Dec 11 2014 18:51
no, I think appear.in only supports 8
Brian E. Granger
@ellisonbg
Dec 11 2014 18:51
no only 8 :)
Thomas Kluyver
@takluyver
Dec 11 2014 18:51
create our own video conferencing system in the notebook!
Fernando Perez
@fperez
Dec 11 2014 18:51
ah
Min RK
@minrk
Dec 11 2014 18:51
I don't think we need more than 10 people in any meeting
Fernando Perez
@fperez
Dec 11 2014 18:51
yup!
Kyle Kelley
@rgbkrk
Dec 11 2014 18:51
That was me testing by hand months ago though
Brian E. Granger
@ellisonbg
Dec 11 2014 18:51
Yes,
Min RK
@minrk
Dec 11 2014 18:51
we just need smaller meetings
Fernando Perez
@fperez
Dec 11 2014 18:51
good pt
maybe that 10 limit is a good thing...
Kyle Kelley
@rgbkrk
Dec 11 2014 18:52

we just need smaller meetings

Yes.

Brian E. Granger
@ellisonbg
Dec 11 2014 18:52
Yeah probably
Damian Avila
@damianavila
Dec 11 2014 18:52
yep
probably more focused...
Jason Grout
@jasongrout
Dec 11 2014 18:52
it's not working for me to get on.
for maintaining after 3.0, it would be easier if it was split out before 3.0
Min RK
@minrk
Dec 11 2014 18:53
that's true of all components of IPython
Fernando Perez
@fperez
Dec 11 2014 18:53
Scott or Damian, can I bug you to follow on YT for a bit? Jason is trying to talk...
Somehow it wasn't enough for just Jess to drop
Matthias Bussonnier
@Carreau
Dec 11 2014 18:53
I'm goin to drop soon .
Fernando Perez
@fperez
Dec 11 2014 18:53
ok
thx~!
Damian Avila
@damianavila
Dec 11 2014 18:54
I drop out...
Jason Grout
@jasongrout
Dec 11 2014 18:54
re horses and ponies: "That's just crazy, we should just think, sleep on it, and move on for now."
Fernando Perez
@fperez
Dec 11 2014 18:54
Jason, go ahead.
Matthias Bussonnier
@Carreau
Dec 11 2014 18:54
:-)
Damian Avila
@damianavila
Dec 11 2014 18:54
maybe 9 is too much for the hangout too...
Scott Sanderson
@ssanderson
Dec 11 2014 18:54
please don't do that...
you probably want to be clear that both the code location and the API might change
Matthias Bussonnier
@Carreau
Dec 11 2014 19:04
I'll drop in 1 minutes. See you soon.
Damian Avila
@damianavila
Dec 11 2014 19:04
See you... good night!!
We need to write JS with JS manners... not pretending writing a pythonic JS...
Jason Grout
@jasongrout
Dec 11 2014 19:08
is @fperez talking about docs or the language?
but es6 is much more pythonic... :)
Kyle Kelley
@rgbkrk
Dec 11 2014 19:11
maybe put your vote next to the one you want in hackpad
Jason Grout
@jasongrout
Dec 11 2014 19:11
By the way, totally off topic: sage math cloud is now opensource: https://github.com/sagemath/cloud
GPLv3
Kyle Kelley
@rgbkrk
Dec 11 2014 19:11
nice
GPLv3?!?!?!
Scott Sanderson
@ssanderson
Dec 11 2014 19:11
@jasongrout cool
Kyle Kelley
@rgbkrk
Dec 11 2014 19:11
Can I read that code and not taint other code?
Jason Grout
@jasongrout
Dec 11 2014 19:12
your eyeballs are then derivative works...not allowed.
Kyle Kelley
@rgbkrk
Dec 11 2014 19:12
(oooohhh coffeescript, don't tell anyone but I enjoy coffeescript)
GPL was forced because of requirements of the grants funding its development, IIRC
Scott Sanderson
@ssanderson
Dec 11 2014 19:13
@rgbkrk all our javascript at Q is coffeescript
Jason Grout
@jasongrout
Dec 11 2014 19:13
though he did BSD-license things like the differential syncing.
Kyle Kelley
@rgbkrk
Dec 11 2014 19:13
ah, cool
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:13
I don’t know how far behind the youtube is, but I am also in favor of the first one
Kyle Kelley
@rgbkrk
Dec 11 2014 19:13
@ssanderson that's awesome
Jason Grout
@jasongrout
Dec 11 2014 19:14
Also, way off topic, Chris Colbert is adopting TypeScript for PhosphorJS

"Question: Why is SMC open source?

Answer: Two of the four NSF grants that very substantially supported
SMC development had explicit open source requirements. " - William Stein

Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:36
Is there a way to validate the contents of a trait? I know there’s _traitname_changed, but that doesn’t apply to default values, correct?
Thomas Kluyver
@takluyver
Dec 11 2014 19:36
isn't there a validate() method?
oh, right, but you have to get the trait instance
I forget if there's a public API for that
@minrk might know
Min RK
@minrk
Dec 11 2014 19:37
I think creating a new traitlet is appropriate
it's fine to have special use traitelets defined in subpackages
I don't think it's possible to add a validate function to be called, though you can raise TraitError in on_trait_change
Scott Sanderson
@ssanderson
Dec 11 2014 19:38
it might be nice to have a Predicate trait or somesuch that takes a validation function to its constructor
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:38
Ok, cool, thanks
Min RK
@minrk
Dec 11 2014 19:39
That's what we do most places - in _[trait]_changed, check the value, and either assign a new value, or raise a TraitError
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:39
Yeah
Min RK
@minrk
Dec 11 2014 19:39
but it would be perfectly fine to do that in a new traitlet with a validate method as well
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:39
Does _[trait]_changed get called when the initial value is set?
No, if you want true validation of any possible value, a new traitlet is the only option
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:40
Ok
Min RK
@minrk
Dec 11 2014 19:40
but on_changed will validate any custom value (it wouldn't catch the case in your PR)
Jessica B. Hamrick
@jhamrick
Dec 11 2014 19:40
Right
Kyle Kelley
@rgbkrk
Dec 11 2014 19:47
@jdfreder I'll start reviewing ipython/ipython#7116 now that we've made a decision
@jdfreder I've been used to using style like this for objects: https://github.com/kenperkins/coreos-cluster/blob/master/index.js#L6-L19
Jonathan Frederic
@jdfreder
Dec 11 2014 20:08
@rgbkrk still here?
Kyle Kelley
@rgbkrk
Dec 11 2014 20:09
yeah
looking through your PR
Jonathan Frederic
@jdfreder
Dec 11 2014 20:09
Thanks I saw that
I have a question
Kyle Kelley
@rgbkrk
Dec 11 2014 20:09
Pretending to learn how the JavaScript works in the notebook
You're the expert, I'd just like to learn more and am trying to take review on.
Feel free to ask away and tell me I'm doing it all wrong.
Jonathan Frederic
@jdfreder
Dec 11 2014 20:10
What do you think is better, integer, which is more descriptive of the actual value, or number, which is the actual type?
Kyle Kelley
@rgbkrk
Dec 11 2014 20:11
I guess the real question is how the generated JSDocs handle it
Thomas Kluyver
@takluyver
Dec 11 2014 20:11
would_be_an_integer_if_language_understood_that ;-)
Kyle Kelley
@rgbkrk
Dec 11 2014 20:11
Since Integers don't really exist in JavaScript
yeah
I think @takluyver's is best
Jonathan Frederic
@jdfreder
Dec 11 2014 20:11
hahah
well integer is short for what he typed.
Kyle Kelley
@rgbkrk
Dec 11 2014 20:11
It seems like people use integer
Jonathan Frederic
@jdfreder
Dec 11 2014 20:12
And I've been using float otherwise
Kyle Kelley
@rgbkrk
Dec 11 2014 20:12
Alright, we should use integer but interpret as shortening would_be_an_integer_if_language_understood_that
Jonathan Frederic
@jdfreder
Dec 11 2014 20:12
yeah
Kyle Kelley
@rgbkrk
Dec 11 2014 20:12
Otherwise you're going to have write number and then in the comments piece write that it must be an integer
Jonathan Frederic
@jdfreder
Dec 11 2014 20:13
And I think you're right, object > dictionary (> dict, which is Python specific)
(I added the last piece myself)
Kyle Kelley
@rgbkrk
Dec 11 2014 20:14
yeah
I sometimes still say associate array on accident
Guess which language I started with that had that interface. ;)
Jonathan Frederic
@jdfreder
Dec 11 2014 20:15
And also, I think your syntax for object keys would work well. I haven't been using it myself, but it seems like a good way to avoid nestisitis.
Kyle Kelley
@rgbkrk
Dec 11 2014 20:16
yeah
Jonathan Frederic
@jdfreder
Dec 11 2014 20:16
def nestisitis: disease where an excess of indents pushes the content to the right margin, at which point wrapping causes the text to be reduced to one word per line.
Kyle Kelley
@rgbkrk
Dec 11 2014 20:16
It's been clean
Oh em gee, I'm done scanning through
Jonathan Frederic
@jdfreder
Dec 11 2014 20:18
Ok thanks, when I address the comments I'll ping back.
Kyle Kelley
@rgbkrk
Dec 11 2014 20:18
I felt bad telling you to uppercase some of the comment starts
For the sake of consistency
I'll just write a summary
Jonathan Frederic
@jdfreder
Dec 11 2014 20:18
No don't feel bad
I agree
Most of them were inconsistent to start
Kyle Kelley
@rgbkrk
Dec 11 2014 20:20
yeah, I'm not nitpicking on you just that its there and we can clean it up
:)
Cool, after you fix those little pieces I'll merge
Jonathan Frederic
@jdfreder
Dec 11 2014 20:20
Honestly I haven't looked at the grammar yet, but it makes sense to do so now. Especially since we will be building docs from this.
Min RK
@minrk
Dec 11 2014 20:22
We might hold #7116 if it's going to conflict with #6700, which has been open for a long time, and is close to ready
Kyle Kelley
@rgbkrk
Dec 11 2014 20:23
ah, k
Thanks for letting me know before I made a mess @minrk !
Min RK
@minrk
Dec 11 2014 20:37
@ellisonbg latex question?
Kyle Kelley
@rgbkrk
Dec 11 2014 20:39
\begin{question}
Min RK
@minrk
Dec 11 2014 20:53
ha
Jonathan Frederic
@jdfreder
Dec 11 2014 21:27
\end{question}
Parser error:
Unexpected end of question.  Perhaps a question was not provided.
Thomas Kluyver
@takluyver
Dec 11 2014 21:28
Emergency stop!
Jonathan Frederic
@jdfreder
Dec 11 2014 21:28
lol
Brian E. Granger
@ellisonbg
Dec 11 2014 21:33
@minrk sure
was on the phone with fernando
but free now
although given the syntax of @rgbkrk and @jdfreder it looks like your question is "ha"
Jonathan Frederic
@jdfreder
Dec 11 2014 21:36
@ellisonbg the question macro is a little more powerful than that. It uses a regex to check if the text is followed by a ?, but since there wasn't a ? it failed to parse. @minrk 's question==None
Min RK
@minrk
Dec 11 2014 21:57
@ellisonbg should display.Math make inline or display math? ($$ vs $)
Brian E. Granger
@ellisonbg
Dec 11 2014 22:34
Hmm, what does it do now @minrk ?
Min RK
@minrk
Dec 11 2014 22:34
display
Brian E. Granger
@ellisonbg
Dec 11 2014 22:34
That was my first thought just now
Min RK
@minrk
Dec 11 2014 22:34
okay
Brian E. Granger
@ellisonbg
Dec 11 2014 22:34
Why is this coming up?
But I don't have super strong feelings -
Min RK
@minrk
Dec 11 2014 22:35
#7184
in latextools, there's latex_to_png
Brian E. Granger
@ellisonbg
Dec 11 2014 22:35
Does this relate to how math is indented in the notebook
Ahhh, got it
Min RK
@minrk
Dec 11 2014 22:35
in the qtconsole, it's called by stripping $ and then passing wrap=True
which is certainly wrong, because it breaks cases like 'where $a=5$'
Brian E. Granger
@ellisonbg
Dec 11 2014 22:36
Yes
Min RK
@minrk
Dec 11 2014 22:37
but mpl mathtext doesn't support display math, only inline, so when I removed that, it started failing
In that PR, I hardcode text.replace('$$', '$') in mpl_tex2png
Brian E. Granger
@ellisonbg
Dec 11 2014 22:38
So the main question is how should the latextools stuff handle all of this - rather than how display.Math works
Min RK
@minrk
Dec 11 2014 22:38
partly
I was thinking earlier than inline might be preferable, but now I think display is better
Brian E. Granger
@ellisonbg
Dec 11 2014 22:40
I think is my preference by a little bit
Min RK
@minrk
Dec 11 2014 22:40
that PR should fix various things, including rendering latex in the qtconsole, which just doesn't work right now in a variety of ways
One can always use Latex('$inline$') to get inline math, so it's not of any great consequence what we do in Math.
Brian E. Granger
@ellisonbg
Dec 11 2014 22:41
are their remaining questions to figure out for #7184 ?
Min RK
@minrk
Dec 11 2014 23:15
Don't think so, should be okay.