Re: [Mingw-users] Shared libstdc++ problem testcase [Was: Creating ashared libstdc++]

[Greg C. <chi...@co...>]

On 2006-12-5 5:03 UTC, Danny Smith wrote:
> 
> Greg Chicares
> Sent: Tuesday, 5 December 2006 4:47 p.m.
>>
>> The application failed to initialize properly (0xc0000005). \
>> Click on OK to terminate the application.
> 
> I thought there was a follow message after the recipe which warned that
> a C++ frontend patch was needed to avoid problems with auto-import of TI
> objects from .rdata sections. Namely that the loader doesn't like being
> lied to about .rdata.  No? Well  there should have been 

"Zum Donnerwetter!"--there was, on 2006-11-16 20:50 UTC, when you wrote:

|> (Translation: The application could not be initialized (...).
[...]
|
| Bugger! Pshaw! Thunderbolts! (My year 5 teacher in primary school used
| to say something like that in German)
|
| I'm guessing that is another one of those .rdata typeinfo thingies.
| Are you willing to continue testing with gcc-3.4.5  or would you prefer
| to try
| gcc-4.2.0?  No warranties, but it does have that new-car smell about it.

I'm stuck at gcc-3.4.4 due to this problem with gcc-3.4.5:

  http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12477
| warning: inline function [...] is declared as dllimport: attribute ignored

because I really want to use '-W -Werror'. The answer to that problem is
'-Wno-attributes', which was added only on gcc-4.x .

So I guess I'll have to wait for gcc-4.2.0 .