lkml.org 
[lkml]   [1997]   [Jan]   [21]   [last100]   RSS Feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
/
From"Mauro Condarelli" <>
DateTue, 21 Jan 1997 16:56:13 +0000
SubjectRe: 2.1.21 patches for a clean compile
> On Thu, 16 Jan 1997, Christian Kirsch wrote:
> >
> > Alan Cox wrote:
> > > 
> > > I'm much more concerned about having _correct_ drivers than ones that happen
> > > to compile. 
> > 
> > This is a very subtle and fine point that I'm probably not intelligent enough to understand: A
> > driver that fails to compile but still works. Mut be something virtual ;-)
> 
> Alan is very good in saying 'all' with an absolute minimum of words. ;-)
> 
>   A driver that has been proven to be _correct_ implies beeing successfully 
>   compiled. However, a driver that just successfully compiles does _not_ 
>   imply beeing correct (though it may).
> 
> So, nothing 'virtual', just pure logic.
> 
> Hans
> <lermen@fgan.de>
Sorry.
You missed the point completely.
What Alan was trying to say (unsuccessfully, since the previous 
messagesages) is (if i'm not *badly*wrong :) :):
Compiler warnings are not a concern of mine, as long as they are well 
understood.

This holds true (IMHO) because the compiler keeps "on the safe side", 
so it issues warnings also in situations where they can not lead to 
real malfunction. Changing the sources to stop gcc from complaining 
(putting a cast somewhere, for instance) can be useless and even 
dangerous, since it removes a reminder about something that should be 
cared for.

No one was speaking about non-compiling code (of course).

Just my $.02
Mauro

\
 
 \ /
  Last update: 2005-03-22 12:38    [from the cache]
©2003-2008