fix statement-end detection in qtconsole

[stevengj]

Currently, the qtconsole front-end is broken because it is using a Python heuristic for detecting whether the statement is ended (i.e. whether starts a new line in the same cell or completes the cell and requests execution).

@fperez mentioned this issue, but I forget where he said it had to be fixed. (Somewhere in the front end, though, not in the back-end kernel.)

[minrk]

The QtConsole uses an InputSplitter to determine whether the current input block is complete and should imply execute here.

The long-term fix (hopefully this Fall) is to add a message for this, so frontends can query the kernel for this information rather than all frontends having to know about all kernel languages.

The short-term workarounds are:

• use ctrl-enter to skip the complete-input check
• monkeypatch is_complete to always return False, which would force shift-enter to execute (like the notebook)

We can probably add a simple switch in IPython that lets you skip input-checking with a flag rather than a patch to the frontend.

[StefanKarpinski]

Fall seems awfully far away – is there some dependency or release policy that prevents this being done sooner?

[Carreau]

Fall means it will probably be implemented by Fall, just a question of time.
That would mean that the code will not be into stable before December
(releases every 6 month). An now we are in feature freeze for 1.0 (due in
2weeks)

Le dimanche 28 juillet 2013, Stefan Karpinski a écrit :

Fall seems awfully far away – is there some dependency or release policy
that prevents this being done sooner?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/22#issuecomment-21687898
.

[minrk]

It might be next week if you live on master, I just think it will miss the 1.0 cutoff since it's technically a new feature.

[StefanKarpinski]

Fair enough. A feature freeze is a feature freeze.

[minrk]

For instance... ipython/ipython#3814

[Carreau]

We'll try to add the ability to disable autoexec for 1.0 (we really love you guys), but changing the message spec for 1.0 might be a lot.

[StefanKarpinski]

That seems like a reasonable compromise – thanks!

[stevengj]

Closing for now, since this is a reasonable interim solution pending spec updates.

[malmaud]

@Carreau @minrk Any updates on this?
I'm trying to get IJulia qtconsole up to snuff: customizing the new FrontendWidget.lexer_class option is giving me good Julia syntax highlighting, but we still need statement-end detection and autoindenting.

[minrk]

No, there has been no progress on extending the statement parsing in frontends. A proposal has been made to add a message for querying the kernel, since that's where this knowledge actually resides.