The shim is a tiny bit more complicated now, since it's not just moving IPython.frontend.* to IPython.* anymore, as IPython.frontend.html.notebook became just IPython.html. A result of the extra shim is that there is one old import that doesn't work:

from IPython.frontend.html import notebook

but

from IPython.frontend.html.notebook[.submod] import whatever

works just fine.

It seems to be an artifact of sys.modules['IPython.frontend.html.notebook'] already existing as a module, but not being a submodule of what IPython.frontend.html would point to.

You don't plan to have html-based console? If so, I think IPython.html would be not the bast place for notebook.

I'd like to have an html console, but I think it will be part and parcel of
the notebook interface, not something separate. In that sense, a common
'html' directory makes sense.

On 20 June 2013 16:48, Takafumi Arakaki notifications@github.com wrote:

You don't plan to have html-based console? If so, I think IPython.htmlwould be not the bast place for notebook.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/3450#issuecomment-19762226
.

We've learned from the entangling of HTML things and static file dirs, that even if/when we add an HTML console, we will probably still only have one main webapp, and just add to it. Console-related bits will get their own service-related subdirs, etc., but it wouldn't be an independent entity.

It looks like this one has a conflict again...

rebased

@ellisonbg @fperez I thought this was supposed to merge before too many more PRs start coming in. I've had to rebase it three times already.

Will try to merge today at scipy sprints

Sent from my iPhone

On Jun 27, 2013, at 1:15 PM, Min RK notifications@github.com wrote:

@ellisonbg @fperez I thought this was supposed to merge before too many more PRs start coming in. I've had to rebase it three times already.

—
Reply to this email directly or view it on GitHub.

@ivanov and I have run the tests successfully, although we had to resolve some issues with stale bytecode files.

my tests run ok too... great