[prev in list] [next in list] [prev in thread] [next in thread] 

List:       hpux-cxx-dev
Subject:    CXX-DEV: =?iso-8859-1?Q?RE=A0=3A_CXX-DEV=3A_abort_on_malloc_failure?=
From:       "Maucci, Cyrille" <cyrille.maucci () hp ! com>
Date:       2005-04-14 9:13:43
Message-ID: 248289F43319004F8F13A4501D47D15801309637 () idaexc02 ! emea ! cpqcorp ! net
[Download RAW message or body]

ooopsss reply all too fast...

________________________________

De: Verney, Remi
Date: jeu. 14/04/2005 11:11
@: Maucci, Cyrille
Cc: cxx-dev@cxx.cup.hp.com
Objet : Re: CXX-DEV: abort on malloc failure



Hi man,
Tu as des soucis avec malloc() ? Si tu veux un coup de main... Je me
fais unpeu chier ici tout seul au B1N2....
BTW, des outils comme electric fence marchent assez bien, ca m'a
economise pas mal de migraines... http://perens.com/FreeSoftware/
Ca detecte les allocations qui foirent, les allocations de 0 byte, et
les debordements de buffers alloues lors de leur utilisation, qui des
fois pourrissent les structures de donnes du tas et qui font planter
malloc ou free ensuite... Ca peut bien etre ca qui t'arrive, j'ai deja
eu ca.

   Remi

Maucci, Cyrille wrote:

>All,
>
>
>
>I'd like something to be confirmed here.
>
>I am almost sure to have already learnt on this list that on malloc
failures,
>in some cases (maxdsiz or ulimit reached?, no more memory available on the
>system?), the executable could abort with SIGABRT or SIGSEGV.
>
>I am not sure if this is right or wrong, and my attempts to reproduce such a
>behavior was unsuccessful: for example, I wrote a program that does malloc a
>lot of big chunks of mem, but malloc simply fails without the executable
being
>aborted.
>
>
>
>Can anybody confirm if SIGABRT or other signals can happen or not? And if
yes,
>what are exactly the different cases for which that can occur?
>
>
>
>Many thanks
>
>Cyrille
> _________________________________________________________________
> To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
>    with the message UNSUBSCRIBE cxx-dev
> _________________________________________________________________
>
>
>
>
 _________________________________________________________________
 To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
    with the message UNSUBSCRIBE cxx-dev
 _________________________________________________________________
[prev in list] [next in list] [prev in thread] [next in thread]