Shift-Enter to run a cell is not cached

Description

What steps will reproduce the problem?

If you have two consecutive code cells (starting with a #%%) and you try to run them both by pressing Shift-Enter twice, only the first cell will be run.

This happens when the first cell takes some time to run, and it’s still running when you press the second shift-enter. The second shift-enter is ignored.

In version 3.X it works fine, running the two cells one after the other.

Versions

• Spyder version: 4.0.0b3
• Python version: 3.7.3
• Qt version: 5.12.3
• PyQt5 version: 5.12.2
• Operating System: Windows 10

Dependencies

pygments >=2.0    :  2.4.2 (OK)
qdarkstyle >=2.6.4:  2.7 (OK)
sphinx >=0.6.6    :  2.1.0 (OK)
pyls >=0.27.0     :  0.27.0 (OK)
nbconvert >=4.0   :  5.5.0 (OK)
pandas >=0.13.1   :  0.24.2 (OK)
numpy >=1.7       :  1.16.4 (OK)
sympy >=0.7.3     :  None (NOK)
cython >=0.21     :  0.29.10 (OK)
qtconsole >=4.5.0 :  4.5.1 (OK)
IPython >=4.0     :  7.5.0 (OK)
matplotlib >=2.0.0:  3.1.0 (OK)
pylint >=0.25     :  2.3.1 (OK)

[impact27]

Yeah that’s great idea. Definitely the way to go. The queue would then be a simple list of command strings and the cache would be a dictionary of code and hashes. We can have Spyder do all the code caching separately when it adds to the queue.

We could add an optional parameter to the runcell and runfile like cache=None which would be a little cleaner than adding it to the cellname and then having to parse it. Then we would be be able to cache whole files for runcell as well(the imported modules could still change but we have to stop somewhere).

I think this is the correct way of doing this. I modified runfile in https://github.com/spyder-ide/spyder/pull/10115 to get the code from the editor, so we could even cache runfile. When this PR is merged, a cache parameter can be added to _handle_get_file_code and handle_run_cell.

[bcolsen]

One way I can think of would be to add something to the cell name: if we want to run the cell “do stuff”, and the kernel is busy, we could cache the cell content and give it a name like “do Stuff @ cache:1234”. When the line is executed, the run_cell function will ask the frontend for the code for cell “do Stuff @ cache:1234”, and the frontend could pop the cache. This way we don’t need to modify the kernel.

Yeah that’s great idea. Definitely the way to go. The queue would then be a simple list of command strings and the cache would be a dictionary of code and hashes. We can have Spyder do all the code caching separately when it adds to the queue.

We could add an optional parameter to the runcell and runfile like cache=None which would be a little cleaner than adding it to the cellname and then having to parse it. Then we would be be able to cache whole files for runcell as well(the imported modules could still change but we have to stop somewhere).