A BEGIN_SRC block can specify which mime type to display by enabling the
user to manually set the priority order of mime types.
e.g.: In the following, :text/plain will be used before considering :text/html
#+BEGIN_SRC jupyter-python :display plain html
When inserting continuation prompts, extra text deletion entries where being
added in `buffer-undo-list` which caused yanked text to not be undone fully
since it would add in those entries before undoing the yank.
If both plain text and markdown are provided, use markdown. IJulia provides
both representations when, e.g. asking for help on a symbol using ? at the
REPL.
The modification-hooks added by `org-format-latex` cause the overlay to be
deleted whenever it is modified and this will also happen during text property
changes which we want to avoid. Since the buffer is read-only anyways, there is
no need for this hook.
`jupyter-org-with-src-block-client` is used in the `completion-at-point`
interface and `with-current-buffer` is slow so there should be a slight speed
up here.
The "current input cell" is intended to be the last cell in the REPL buffer so
go to `point-max` before calling `jupyter-repl-cell-beginning-position`.
* Add `jupyter-repl-syntax-propertize-function`. This allows packages like
`rainbow-delimiters-mode` to work regardless of what is printed as output of
a REPL cell by setting the syntax-table property on parenthesis in REPL
output.
* Fix an issue with the Julia REPL which would cause `syntax-ppss` to
report the wrong parenthesis depth due to the `]` character of the REPL
package mode. Thus interfering with how `rainbow-delimiters-mode` works.
* Ensure that the kernel language's syntax table is used when fontifying and
adding syntax properties.
There is no need to reset the prompt when it isn't displaying the cell count.
This also ensures that a kernel that shows a prompt other than the input prompt
doesn't get overwritten.
