No need to reload session (I doubt creating a new folder create a
session), and does it once the creation drive promise return which make
the new directory appear in the listing for slow remote-backend (like
google drive)
Mitigate 7486, could probably be improved, but
treat text as pure text, but assume that when received a dome
element it actually come from a secure source.
Sanitizing all we get, including Dom Element break json-metadata editor
and a few others.
fixes#7355
Also limit the number of query "jump", and use variables.
In particular there was a container width discontinuity at 768 pixel
where below 768 container width is 100% and "only" 720px on wider
screen, making a jump.
So now Css media query have 3 breakpoints:
- <540 px with (aka -no name-)
- colapsed menubar
- vbox for prompt
- 100% width container
- 540 -> 767 (aka -xs- )
- expand menubar
- hbox fro prompt
- 767 -> 991 (aka -sm-)
- show save widget
- fix width (768px)
- 991+ (aka -md-)
- extra save widget info (date time)
- fix with of 991 px.
- 1200+ (aka -lg-)
- widht of 1200px.
of course we've set soem variable wrongly so this is still kinda-a-hack
as we shoudl have pushed the xs-sm limit lower instead of pushing the
nothing-xs limit up but I don't want to "fix it" now.
- kernel_selector.set_kernel validates selection and triggers 'spec_changed.Kernel'. It does not start the session anymore.
- notebook calls kernel_selector.set_kernel when:
- kernelspec is in notebook metadata
- session is loaded (e.g. no kernelspec metadata)
- notebook starts session, loads metadata on spec_changed.kernel
The only case where starting the session is not triggered by spec_changed is on notebook load with no kernel metadata
closes#7409
Mostly a symptome of too many nested div that do different things.
don't try to both have an end space inside and outside the 'document'
area. And don't try to also get the things to be 100viewport height by
hacking around and make them smaller;
- Adds a `GenericCheckpointMixin` as a helper for implementing the two
boundary-traversing Checkpoint API methods, `create_checkpoint` and
`restore_checkpoint`.
- `GenericFileCheckpointManager` is implemented as a subclass of
`FileCheckpointManager` using `GenericCheckpointMixin`. Note that
this is the safe subtyping relationship because of method
signature *contra*variance: `FileCheckpointManager` accepts
`FileContentsManager` in its method signatures type, whereas
`GenericFileCheckpointManager` accepts any `ContentsManager`.
- Moved Checkpoint-related classes to their own files.
On `create_checkpoint` and `restore_checkpoint`, pass a path and a
`ContentsManager` to `CheckpointManager` instead of an already-loaded
model.
The `CheckpointManager` base class provides a correct implementation of
these methods that's generic across any ContentsManager, but subclasses
are free to specialize when the storage backend of `ContentsManager` is
shared.
multiplex on a 'channel' key in message,
rather than separate websockets.
Unlike zmq, there aren't different message patterns that require the channels to be separate.
Reduces FD count by factor of 3 and connection complexity in js.
rather than in a dict, which doesn't work without reserializing *all* stored widget states (of all notebooks), since localStorage only supports storing strings.
When a users uses the rename dialog in the text editor, we call
_set_mode_for_model on the returned model. The expected model to be
returned from rename has no value for mimetype, which causes us to pass
`undefined` to `CodeMirror.findModeByMIME`, which returns `undefined`
back. We then try access an attribute of the (undefined) return value,
causing an error.
This now properly checks whether the `mimetype` attribute is set on the
input model.
Changed the public checkpoint API to:
- `create_checkpoint(self, nb, path)`
- `get_checkpoint_content(self, checkpoint_id, path)`
- `rename_checkpoint(self, checkpoint_id, old_path, new_path)`
- `delete_checkpoint(self, checkpoint_id, path)`
- `list_checkpoints(self, path)`
All paths in the above are API-style paths, and the `nb` argument to
`create_checkpoint` is a dictionary suitable for passing to
`nbformat.write`.
The new `get_checkpoint_content` method returns an unvalidated notebook
content dictionary. It is used by `ContentManager.restore_checkpoint`
to load content to be written via `save`.
Ensures kernel.js is always loaded.
It wasn't being loaded when creating a new notebook with a particular kernel because `change_kernel` wasn't being called. Only the `spec_changed` event is triggered by all the various ways a kernel can be loaded,
so load kernel stuff on that event.
- Add a `CheckpointManager` base class and infrastructure for creating a
`checkpoint_manager` instance attribute on `ContentsManager`.
- Provide default implementations of `delete` and `rename` in the base
`ContentsManager` class. `ContentsManager` subclasses are now
required to implement `delete_file` and `rename_file`. These methods
no longer need to manage checkpoints.
- Move checkpoint-related functionality from `FileContentsManager` to a
dedicated `FileCheckpointManager` subclass.
- Move shared filesystem interaction logic into `FileManagerMixin`
used by both `FileContentsManager` and `FileCheckpointManager`.
- Minor tweaks to ContentsManager tests to get methods from the right
object.
The purpose of this change is to provide an API for users to replace
just the checkpoint logic associated with a particular
`ContentsManager`. In particular, this change makes it possible to
easily support remote storage of checkpoints while otherwise retaining
normal filesystem interactions.
closes#7303.
Also change the arrow that indicate submenu for a fontawesoem icon,
instead of the ugly hack that show only the left border of a rectangle element
which with is twice the one of its border to get a triangle.
Juergen Hasch
Juergen Hasch
Matthias Bussonnier
Min RK
Thomas Kluyver
Andrew Dawes
Mathieu
Nicholas Bollweg
Brian E. Granger
Jason Grout
Scott Sanderson
Jonathan Frederic
dongweiming
Jonathan Frederic