Re: Problem with AC_CHECK_TYPE on MacOS X

[Reuben Thomas]

On Tue, 31 Mar 2009, Eric Blake wrote:

> But the mere fact that we are trying to #define size_t, which then
> conflicts with a different set of header includes

Sorry, I wasn't clear. When I looked at the config.log, attached, I looked 
first at the check for size_t, as you suggested, so see why the system's 
size_t was not being accepted. As you can see, on this system, the check for 
size_t fails to compile because of parsing errors before reaching the code; 
at this point, size_t has not been redefined (that's in a later test, after 
this one has already failed).

> It would be really nice to see the config.log output of the AC_TYPE_SIZE_T 
> failure,

(attached)

> as well as the preprocessor output of the conftest.c test that shows how
> the system headers are inconsistent under -std=c89.

How would I generate this easily?

Sorry, I'm not trying to withhold anything, it's just that I'm a proxy in 
this case; Nelson Beebe has the machines doing the building, and he seems to 
be too busy to report these problems himself, so I'm having to get the 
details from him, and obviously that involves trying to keep my requests 
trivially simple for him.

--
http://rrt.sc3d.org/ | plagiarism, n.  the mind burgles

config.log.zile-2.3.5.r4-fuji.bz2
Description: Binary data