Re: [Bug-apl] edif update / ⎕IO is 0

[Hans-Peter Sorge]

Hello Chris,

==== remark

I am still puzzled - or better I have no clue ....

 -- The intermediate output is being written to the terminal by stdout /
stderr.

 -- within an apl session I can scroll back and forth. The additional
lines from the above mentioned output

   might interfere with with apl internal scrolling. A wild guess.


==== new case
There is an other newly discovered quirk in a new apl session:

      'libedif2.so' ⎕fx 'edif2'
edif2
      )load times
SAVED 2018-08-16 22:46:27 (GMT+2)
      ^[[A^[[A^[[A
⍝ The cursor gets stuck before ^ instead of cursor moves I get control
sequences
⍝ An ENTER fixes this. 


==== more findings
Next: Lots of "again" lines ( about 71, shortened )

address@hidden ~]$ apl

      )load times
SAVED 2018-08-16 22:46:27 (GMT+2)
      'libedif2.so' ⎕fx 'edif2'
edif2

⍝ so far so good     
)fns                                                                            
                                                                                
                                          

Minutes     edif2   times


⍝ start editing

       edif2 'times'

⍝ output from edif2:
again
........... several lines
again


⍝content of times:
∇times[⎕]∇                                                                      
                                                                                
                                          


    ∇
[0]   times
[1]  ## 02:00

........... similar lines

[42] ## 02:15
[43] 
[44] 
[45] 
    ∇

⍝ output from edif2:

       edif2 'times'
again
........... several lines
again


⍝ output from edif2:

      'emacs' edif2 'times'
again
........... several lines
again


⍝ and then - bingo

       'emacs' edif2 'xxxx' 
Speicherzugriffsfehler (Speicherabzug geschrieben)

---- looks like the memory corruption is related to the additional, non
apl output.

The "again" output is related to the "mq_send" command.
When I edit the function a second/third/... time, the mq_send section
gets into an infinite loop. 
When I reduce the number of lines in above times function to 32 the
"again" output disappears.

I hope this helps somehow to get closer to the root cause.

Best Regards
Hans-Peter



Am 23.08.2018 um 06:23 schrieb Chris Moller:
> Yeah, I've noticed that APL gets unhappy when stdout/stderr interferes
> with its pty.  I tried to pull down a trial version of slickedit just
> to see what was going on, but stopped when it asked me for a credit
> card on which to charge 300USD, so I'm only guessing that the output
> you're seeing is stdout or stderr.   I did a little experiment with a
> dummy "editor"  (actually hello.c) just to see what happens when you
> dump a lot of stuff to stderr, but that didn't cause any memory issues
> even when I ran under valgrind.  Something I can try would be to
> close(STDERR_FILENO),and maybe close(STDOUT_FILENO) before
> exec()ing--any  editor edif2 can use will almost certainly use GTK+,
> Qt, or something other than standard IO.  (I'd use the axis form, like
> edif2[1] would suppress stderr, or something similar.  The "corrupted
> double-linked list" message comes from malloc, so I'm guessing
> unallocated memory is being stepped on, but valgrind didn't catch any
> of that so I'm not sure what's causing it.
>
> I'll look into the inotify stuff.
>
> Chris
>
>
>
> On 08/22/18 16:15, Hans-Peter Sorge wrote:
>> Hi Chris,
>>
>> it' a stubborn case (the latest version).
>>
>> I have an other case that fails (vs is script starting slickedit) :
>>
>> #### apl ... session:
>>
>>       'libedif2.so' ⎕fx 'edif2'
>> edif2
>>       'vs' edif2 'xxxx' 
>>       /var/run/user/2000/21616/xxxx.apl     ### output from scrip vs
>>
>> ### output from slickedit ------------->
>>
>> /opt/slickedit-pro2015/bin/vs_exe -sc /home/joy/.slickedit -sr
>> /home/joy/.slickedit /var/run/user/2000/21616/xxxx.apl    ### slickedit
>>
>> SlickEdit: An instance of SlickEdit is already being displayed on this X
>> server. The existing instance is being activated.
>>
>> If you want to bring up a new copy of SlickEdit, use +new option.
>>
>> You can turn off this message by setting
>> VSLICKXNOPLUSNEWMSG=1                                       
>>
>> ### output from slickedit -------------<
>>
>>
>> ### this is intermittent. might happen after several editor calls
>>
>>       'vs' edif2 'xxxx' 
>> corrupted double-linked list
>> Abgebrochen (Speicherabzug geschrieben)
>>
>> ### session end
>>
>>
>> The message sent from slickedit seems to corrupt the memory.
>>
>> The problem happened in previous version too (w/o pthread_mutex_......)
>>
>>
>