[CLN-list] future GCC 4.3 and modules.h

[Ralf Wildenhues]

Hello Richard,

* Richard B. Kreckel wrote on Sun, Jan 06, 2008 at 03:08:15AM CET:
> I wrote:
>> So, CL_REQUIRE can't call the global ctors directly, because it doesn't 
>> know what the symbol name is. Only CL_PROVIDE and CL_PROVIDE_END do 
>> know that. But, it doesn't appear to be possible to introduce another 
>> layer of indirection in CL_PROVIDE like the attached patch tries to. It 
>> leads to an infinite loop in the global ctors. Right now, I've run out 
>> of ideas.
>
> Well, it turns out that the infinite recursion was due to the braindead  
> way I implemented that other layer of indirection. A plain C function  
> should do the trick.
>
> The attached patch seems to work over here using prerelease GCC  
> 4.3.0-20071130. I would appreciate it if people give it a try and report  
> how it worx 4 them.

That seems to work fine.  Thanks!

Cheers,
Ralf