This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Software Quality Binutils


Hi John,

> But include/ansidecl.h 
> has ATTRIBUTE_RETURNS_NONNULL, which static analysis tools could use
> (but so far as I'm aware, none do).  Nevertheless, I think binutils
> could make more use of these gcc attributes than it currently does.

Agreed.  And I will be happy to review any patches to add such attributes...

>      Or "Remove this conditional structure or edit its code blocks so that 
>      they're not all the same".  How on earth is that a bug ?  It is not
>      even bad coding.
> 
> I think it's flagging things like:
> 
> if (cond)
> {
>   do_this ();
>   do_that ();
> }
> else
> {
>   do_this ();
> }
> 
> ... which it wants changed to:
> 
> do_this ();
> if (cond)
>  do_that ();
> 
> 
> which of-course has the same effect iff evaluating cond does not
> have side effects.

True, but I do not see this as being a bug.  A possible situation
for code quality improvement maybe, but not a bug.

Besides, how do I tell this scanner that actually the code sequence 
is fine as it is and does not need to be changed ?  I am not going 
to change the code just to pacify a scanner.  (I know that you said 
the same thing - I am agreeing with you - so this comment is really 
directed to Christoph).

Cheers
  Nick


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]