Re: Having Trouble with m4 Input (m4 version 1.4.13)

[Tom Browder]

On Tue, Sep 7, 2010 at 12:13, Gary V. Vaughan <address@hidden> wrote:
> Hi Tom,
>
> On 7 Sep 2010, at 23:49, Tom Browder wrote:
>
>> On Tue, Sep 7, 2010 at 11:16, Tom Browder <address@hidden> wrote:
>>> On Tue, Sep 7, 2010 at 10:50, Gary V. Vaughan <address@hidden> wrote:
>> ...
>>> I have to say, though was that my initial excitement [for m4 capabilities] 
>>> was the
>>> avoid-multiple-reading-of-header-files possibility more than anything
>>> else.  And that is still my interest (I see that is one of the
>>
>> Well, duh, I just found "#pragma once" in GNU cpp--the answer for that
>> problem.  I vaguely remember reading about it once but...
>
> Though I haven't checked the docs or the code, I have a pretty strong 
> conviction
> that gcc is actually pretty smart without the pragma too... it remembers any
> headers that are entirely contained in cpp guards (inside the header file),
> and if it encounter another #include of the same file without a state change
> in the guard macro, it ignores the #include and gets on with the rest of the
> file.

Yes, I think that is pretty standard behavior.  The difference I think
is that the pragma results in the file not being opened again, thus
saving a system call and possibly some perfomance gain over a large
system.. The whole thing may be moot for practical purposes though,
but it would be interesting to see some stats.

-Tom