SEGV on here-documents

SEGV on here-documents

[Zoltan Hidvegi]

If a here-document contains a line longer than 255 characters, it causes SEGV
(not on all OS, eg. on Solaris it only produces a wrong output, but it causes
SEGV on elf-linux).  I had no time to trace the bug but it may be related to
Peters input patches (in which case it is not reproducible with vanilla
beta10).

Zoltan

Re: SEGV on here-documents

[P.Stephenson]

hzoli@cs.elte.hu wrote:
> If a here-document contains a line longer than 255 characters, it causes SEGV
> (not on all OS, eg. on Solaris it only produces a wrong output, but it causes
> SEGV on elf-linux).  I had no time to trace the bug but it may be related to
> Peters input patches (in which case it is not reproducible with vanilla
> beta10).

Seems to works OK on 2.6-beta10 with my input patches for up to 365
characters on OSF/1, but I'd like to hear more.

-- 
Peter Stephenson <P.Stephenson@swansea.ac.uk>  Tel: +44 1792 205678 extn. 4461
WWW:  http://python.swan.ac.uk/~pypeters/      Fax: +44 1792 295324
Department of Physics, University of Wales, Swansea,
Singleton Park, Swansea, SA2 8PP, U.K.

Re: SEGV on here-documents

[Zoltan Hidvegi]

Peter Stephenson wrote:
> hzoli@cs.elte.hu wrote:
> > If a here-document contains a line longer than 255 characters, it causes SEGV
> > (not on all OS, eg. on Solaris it only produces a wrong output, but it causes
> > SEGV on elf-linux).  I had no time to trace the bug but it may be related to
> > Peters input patches (in which case it is not reproducible with vanilla
> > beta10).
> 
> Seems to works OK on 2.6-beta10 with my input patches for up to 365
> characters on OSF/1, but I'd like to hear more.

On OSF/1 it works for me as well.  It seems that the bad thing happens
somewhere in gethere() in exec.c.  Unfortunately I really have no time to
debug it now, I have to sysadm a large Unix network with 7 different OS and
I'll also continue my studies in the University and give some lectures
and that's a bit too much.

I looked at my rcs and I see that Peter's input patch changes only one line,
hgets() to ingets() so the problem is probably not in input.c.

Zoltan