Re: making undefined behaviour defined
- From: jacob navia <jacob@xxxxxxxxxxxxxxxx>
- Date: Thu, 22 Sep 2005 18:32:01 +0200
I agree with Mr Sossman in considering that integer overflow is a poor example for what you are saying.
In lcc-win32, using a compiler option makes integer overflow TRAP, what (I think) is quite useful.
Better a trap than INCORRECT RESULTS!
Of course only signed integers trap, unsigned do not.
Making mandatory to accept integer overflow would destroy this and make the C language a construct with no hope of getting sensible results and catching programmer errors...
jacob .
- Follow-Ups:
- Re: making undefined behaviour defined
- From: Francis Glassborow
- Re: making undefined behaviour defined
- References:
- making undefined behaviour defined
- From: Francis Glassborow
- making undefined behaviour defined
- Prev by Date: Re: making undefined behaviour defined
- Next by Date: Re: making undefined behaviour defined
- Previous by thread: Re: making undefined behaviour defined
- Next by thread: Re: making undefined behaviour defined
- Index(es):