bug#11868: 24.1.50; [python-mode] Blocking call to accept-process-output with quit inhibited!! (message buffer)

[Stefan Monnier]

>>> with some trail and error, the following patch is suppressing the warning.
>> 
>> Thank you for your attempt at finding a patch that works.  The problem
>> is that the warning is there to indicate that there is a "fundamental"
>> problem, and your patch mostly hides the problem.  Maybe it actually
>> solves it, but maybe not.
>> It depends on whether the surrounding code is prepared to deal with
>> accept-process-output being interrupted by quit.
>> IOW, we have to either find or create a place where the code is prepared
>> to handle an interruption by quit.

> I could understand what you are saying, but have no clue what to do
> further.  It would be great if you can guide me to the next step to
> resolve the issue.

Maybe python.el's maintainer will be better placed for that.
But basically, what is needed is to think first about what should happen
if the user hits C-g in the middle of this accept-process-output: what
can the user mean by that (there might be various cases), what should
the resulting behavior be, and then how to change the code to get
that result.

E.g. a typical problematic issues is the state of the process: the
process will presumably keep running, so what will happen next time
python.el wants to send it more commands?

I don't know enough about python.el's code to be able to answer.


        Stefan