Re: [orbitcpp-list] segmentation faults

[Sam C. <sa...@to...>]

Brian May <ba...@sn...> wrote:
>=20
> Why is it that when ever I make any changes to my code, including just
> adding debug print messages and/or removing the code which caused it
> to crash last time, it always crashes in a completely different way
> :-(
>=20
> Now, for instance, that crash I described before is no longer
> happening, instead it crashes after the function returns.
>=20
> I have even seen errors I have never seen before, like "illegal
> instruction" and "trace trap". Wow!
>=20
> I have this insane feeling that debugging this simply program could be
> rather difficult. :-(

Sounds like you need a malloc debugger like ElectricFence, NJAMD,
dmalloc, etc.

Since I know you're using Debian, I can tell you that Debian has about a
million of them, all of which work differently. Each of them is
effective against different kinds of bugs.
--=20
Sam Couter          |   Internet Engineer   |   http://www.topic.com.au/
sa...@to...    |   tSA Consulting      |
OpenPGP key ID:       DE89C75C,  available on key servers
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C