Re: CLISP EMACS SWANK remote connection crash

C Habs <[email protected]> Wed, 9 Oct 2019 11:05:20 +0000
Newsgroups gmane.lisp.clisp.general
Message-ID <CANe8y8ZnCHXafDs1py4-=KbRAjAf1P8_TLa=fpA0tARM=iBsAQ@mail.gmail.com>
On Wed, 9 Oct 2019 at 07:33, Vladimir Sedach <[email protected]> wrote:

>
> Vladimir Sedach <[email protected]> writes:
>
> > I am going to try to reproduce on 32-bit OpenBSD this week.
>
> Update:
>
> With CLISP 2.49 installed from packages, I get:
> *** - UNIX error 22 (EINVAL): Invalid argument
>
> When trying to start SWANK with slime-20190925.1213 from ELPA.
> Relevant stack frames:
>
> <11/256> #<SYSTEM-FUNCTION SOCKET-SERVER> 4
> <12/252> #<COMPILED-FUNCTION SWANK/BACKEND:CREATE-SOCKET>
>
> So, some kind of error opening the socket. I am not sure whether this
> is related to the segfault.
>
> Instead of debugging, I proceeded to try to compile CLISP from the
> GitLab repository.
>
> The build succeeds, but `make check` fails on test (DEFPARAMETER
> *CALLBACKF* (WITH-C-VAR (X 'UINT->UINT #'CALLBACK) X)) with
> "trampoline: cannot make memory executable"
>
> The ports Makefile for CLISP 2.49 does not have USE_WXNEEDED, so this
> seemed a little suspicious. I added LDFLAGS="-zwxneeded", put the
> build directory on a partition mounted with wxallowed, rebuilt, and
> same thing. So there is something else going on.
>
> I will go back and try to debug the problem with CLISP 2.49 from
> packages. If that does not reproduce the segfault, I have another old
> laptop I can install a 32-bit Linux distribution on, and try
> reproducing there.
>
> --
> Vladimir Sedach
> Software engineering services in Los Angeles https://oneofus.la


Thank you for your time being invested; much appreciated. I received some
comments on the 32bit, suggesting as it was 32bit, it might not be worth
any one's extended time to sort and just move to 64bit platform asap, as
that seemed to work without issue.  I can see the point on moving from
32bit, but I don't agree entirely as there are some very useful 32bit
computers stilll out there and it would be nice to know at which point the
coding caused issues just on 32bit.

I'm afraid my learning-curve on the code base involved is steep :-)

Habs.

_______________________________________________
clisp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/clisp-list