[Patch] Warning for mismatched header guard.

[Matthieu Monrocq]

On Tue, Apr 30, 2013 at 1:38 AM, Richard Trieu <rtrieu at google.com> wrote:

> The attached patch adds a new warning that triggers on mismatched header
> guards.  Typically, a header guard would look like:
>
> #ifndef HEADER_NAME_
> #define HEADER_NAME_
>
> int foo = 42;
>
> #endif
>
> Occasionally, typos will cause the header guard to not function properly
> but still compile, such as:
>
> #ifndef HEADER_NAME_
> #define HEDAER_NAME_
>
> int foo = 42;
>
> #endif
>
> The two problem are:
> 1) The header may be developed and used while only being included once
> with no problems.  Thus, it may take some time before the header guard
> functionality is required.
> 2) When the header is included twice, the error is "redefinition of 'foo'"
> which obscures the actual source of the problem.
>
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>
Hi Richard,

First of all, that certainly is a great idea!

On the same line, during development I have encountered a second "stupid"
pattern:

// A.h
#ifndef A_H
#define A_H

#endif

// B.h
#ifndef A_H
#define A_H

#endif

// C.pp
#include <B.h>
#include <A.h>

// some weird error about B not being defined, WHY ?

It's typically hard to track down, because with the game of bringing
includes in, whilst A is clearly being shadowed "by something" (at least,
clear to a seasoned developer), it's pretty hard to tell WHO is shadowing A.


So I wonder if in the same vein it would make sense to warn where two
different files being included in the same translation unit are relying on
the same header guards.

I have no idea about that false-ratio this could bring in though :/

-- Matthieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130501/bb8d9bb1/attachment.html>