More blocking call to accept-process-output with quit inhibited!!

More blocking call to accept-process-output with quit inhibited!!

[John Shahid]

Hi all,

I using commit '2f55b971b4' and I'm noticing the message mentioned in
the title more often in the echo area.  I don't remember seeing this
message before.  Usually, Emacs will lock up for a few seconds.  I have
to press C-g a few times before Emacs become responsive again.  Is this
related to the tramp/accept-process-output changes that were recently
made ?  I use tramp and usually have a few buffers on remote machines
open.

JS

Re: More blocking call to accept-process-output with quit inhibited!!

[Michael Albinus]

John Shahid <jvshahid@gmail.com> writes:

> Hi all,

Hi John,

> I using commit '2f55b971b4' and I'm noticing the message mentioned in
> the title more often in the echo area.  I don't remember seeing this
> message before.  Usually, Emacs will lock up for a few seconds.  I have
> to press C-g a few times before Emacs become responsive again.  Is this
> related to the tramp/accept-process-output changes that were recently
> made ?  I use tramp and usually have a few buffers on remote machines
> open.

Tramp was hit by this, see bug#34297. This should be fixed in master
with commit b32ac17.

> JS

Best regards, Michael.

Re: More blocking call to accept-process-output with quit inhibited!!

[Michael Albinus]

John Shahid <jvshahid@gmail.com> writes:

Hi John,

>> Tramp was hit by this, see bug#34297. This should be fixed in master
>> with commit b32ac17.
>
> I just tried that commit and tramp stopped working for me (with ssh).
> Tramp just hangs and I have to quit using C-g.

And it did work before? Hmm.

> I attached the tramp log for a working session vs one that hangs.  Let
> me know if I should open a bug, I'm not sure if one already exists to
> track this issue.

Unfortunately, not with tramp-verbose set to 10, which would help in
tramp-accept-process-output. Anyway ...

> 06:36:24.188861 tramp-file-name-handler (1) # Interrupt received in operation (expand-file-name /ssh:dev: nil)
> /home/ubuntu
> 
>  [36;1m2019-02-05 11:36:18 [0;35mip-172-31-64-60 [39min [0;32m~
> [36;1m±[0;32m [0;32m| | [0;32m $[39m 

The remote prompt is full of escape sequences, which is always good for
Tramp trouble. I recommend to remove them, as explained in the Tramp
manual.

If this still doesn't work, you might write a bug report.

Best regards, Michael.

Re: More blocking call to accept-process-output with quit inhibited!!

[John Shahid]

Michael Albinus <michael.albinus@gmx.de> writes:

>> I just tried that commit and tramp stopped working for me (with ssh).
>> Tramp just hangs and I have to quit using C-g.
>
> And it did work before? Hmm.

Yes, it did.  I attached the logs from a working tramp session in the
previous email.

> Unfortunately, not with tramp-verbose set to 10, which would help in
> tramp-accept-process-output. Anyway ...
>
>> 06:36:24.188861 tramp-file-name-handler (1) # Interrupt received in operation (expand-file-name /ssh:dev: nil)
>> /home/ubuntu
>>
>>  [36;1m2019-02-05 11:36:18 [0;35mip-172-31-64-60 [39min [0;32m~
>> [36;1m.[0;32m [0;32m| | [0;32m $[39m
>
> The remote prompt is full of escape sequences, which is always good for
> Tramp trouble. I recommend to remove them, as explained in the Tramp
> manual.
>
> If this still doesn't work, you might write a bug report.

I created bug#34335 with tramp-verbose set to 10.  I also tried to get
rid of the escape characters in the prompt without any luck.

> Best regards, Michael.

Cheers,

JS