Re: [Linux-gpib-general] problems with RH 9.0 and Measurement Computing PCI Board

[Frank M. H. <fm...@us...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 22 May 2003 03:04 am, Neil Phipps wrote:
> Hi Frank,
>
> I've traced my way through things until I found the Segfault ... it
> occurs during the 2nd malloc in yy_create_buffer in ibConFlex.c ... the
> malloc calls for 16386 bytes of memory, and _int_malloc dies ... so I
> changed the buffer size down to 1022 (+2 =3D 1024) and everything seems=
 to
> run OK ... I'm thinking that it maybe a compiler issue ... RH 9.0 runs
> with gcc 3.2, so I tried gcc 3.3, but no success !!!! same issue ...
> which leads me to think it's an implementation issue in my system
> somewhere ... I also tried it with different systems ... the base syste=
m
> is a 1.3GHz Athlon with 128 MB RAM, but I got the same result with a
> twin 2200+ Athlon/2 GB RAM system and a 450 MHz PII with 512 MB RAM ...
> but they are all running RH 9.0 ... this leads me to ask ... What gcc
> version do you use, and what options are invoked ?

I use gcc 2.95.4 and 3.2.3.  Apparently, this kind of problem is due to=20
heap corruption.  You might try playing around with mcheck() to see if yo=
u=20
can find any problems.  I'll try it myself later, but I might have less=20
luck as I've never had the problems you are seeing.  The problem might=20
also be with flex, since ibConfLex.c is auto-generated by flex (I use=20
2.5.4).

>
> Also, once it does run, the simple ibtest example gives the following
> ... any comments ?
>

The errors are normal.  You are trying to do board operations (setting th=
e=20
remote enable and interface clear lines) with a device descriptor.

- --=20
Frank

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE+zT/f5vihyNWuA4URAkeIAJ4rGGzi1N/6XvHJll+zBFyFJQp2oQCdFcBv
o2lbCraVqWs7Hcdvfr6VhGU=3D
=3DbV/o
-----END PGP SIGNATURE-----