Re: [RFC] [PATCH] omnibus header cleanup, certification

[Urs Thuermann <urs@isnogud.escape.de>]

"H. Peter Anvin" <hpa@zytor.com> writes:

> It's also worth noting that there is nothing that it can get confused 
> with and still have a compilable expression.

Well, there are cases, where -1 and (-1) make a difference (see below)
but these are extremely unlikely to appear in the kernel src code.
Thus, I also think these patches are unnecessary.

For example,

    int *p;            and       int *p
    -1[p];                       (-1)[p];

are both valid and compilable code segments, with no undefined or
implementation-defined behavior (as long as p points to an element of
an array other than the first and last), and both code segments do
different things.


> I don't believe the unary-expression patches are necessary.  They are, 
> of course, harmless, except for the fact that my eyes glazed over 
> staring at page after page of these, which very few actual potential (!) 
> bugs (there were a couple, like the iopage+ ones...)

ack.


urs