Repositories of ABCL-specific code? / Kill runaway ABCL function and return to REPL?

[Garrett Dangerfield]

Not sure how ABCL handles overloading, but couldn't you overload the
function that's gone haywire or something that it calls?

On Tue, Jun 11, 2019 at 7:55 PM Alan Ruttenberg <alanruttenberg at gmail.com>
wrote:

>
>
> On Tue, Jun 11, 2019 at 10:35 PM Elliott, Clark <elliott at depaul.edu>
> wrote:
>
>> Dear ABCL experts,
>>
>> I am working on AI system development in LISP. I like using ABCL because
>> of its easy portability to Windows, Mac, and Linux and its direct access to
>> Java. I am running ABCL in emacs shells, often as several processes at
>> once, communicating via sockets. The LISP is working very well indeed.
>> Great job!
>>
>> I've slogged through getting things like sockets and threads
>> running--mostly using apropos and guessing--but it takes a lot of time. I
>> STRONGLY prefer to work from examples of running code.
>>
>> I would dearly love some tips on the following:
>>
>> 1. Are there repositories of ABCL-specific running code, small programs,
>> utility applications and etc. showing the actual running of ABCL
>> implementation-specific functions? Native ABCL networking? ABCL tricks?
>> Examples of running programs using the native threads? Examples of small
>> systems using Java-to-LISP and LISP-to-Java (beyond the given short
>> examples)?
>>
>
> My project LSW2 <https://github.com/alanruttenberg/lsw2> uses ABCL
> extensively. It isn't organized for documentation but browsing through it
> you will find lots of usage of java libraries using JSS. The util directory
> might have some relevant examples. I haven't used native threads - instead
> there is lparallel <https://github.com/lmj/lparallel>
>
>>
>> 2. How can I stop a runaway ABCL LISP function and return to top-level
>> LISP without having to blow away the emacs buffer, and thus my entire
>> current LISP context? In other LISPs I've had some kind of control key
>> sequence, but can't find one for ABCL.
>>
>
> There isn't one because java doesn't allow general interrupts. If a
> process is grinding and never yielding then there's nothing you can do.
> If you are writing the code then you can make sure you don't have long
> loops without every yielding. I tend to throw in a (sleep .01)
> strategically.
> What you can do is have a process run the process you care about, sleeping
> most of the time. When sleeping it will recognize interrupts. The abort
> handler can request a thread be destroyed. Maybe it will be maybe it won't
> but at least you can be returned to the repl. util/geturl.lisp does this. I
> also mucked around with making a unix-like "&" to put a form in the
> background (running in its own thread). That's in util/background.lisp.
> With that you can use (& form) to put form in the background. There's a
> little doc at the top of the file.
>
>
>> E.g.:
>>
>
> Are you using slime? C-c C-c interrupts runaway just fine.
>
>
>>
>> (defun runaway ()
>>   (let ((n 0))
>>     (loop while (< (incf n) 10)
>>        ;; Pretend this ran away. How do I kill it in ABCL REPL?
>>        do
>>          (sleep 1))))
>>
>> Thanks in advance for any help.
>>
>> Best,
>>
>> Clark
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/armedbear-devel/attachments/20190612/1b14ede0/attachment.html>