Hi Doug,

> Not only can the warning be ignored, the warning will no longer occur
> with the binutils from the current trunk HEAD, because
> --enable-auto-import has now become the default for x86 and x64 mingw
> targets.

Yes, I also confirmed this also last night by looking at the latest
binutils from CVS.


> I should mention that the auto import warning is basically telling the
> user that an object was attempting to use symbol foo, and it didn't
> find foo in any of the libraries, but found __imp_foo in one of the
> import libraries and is using that instead.

I see.  That makes sense -- the import library has the __imp_* symbols
but the header file doesn't have the dllimport (as Jon also said
earlier).

One question though: if the libstdc++ header files don't have the
decorations for dllimport, I assume they don't have them for
dllexport.  Then, how do all of the symbols originally get exported
properly to the dll itself?  Is that by using the --export-all-symbols
option during the build?


> With the pseudo-reloc 2
> changes, you have to be doing some really really really odd things to
> run across any issues with enable-auto-import.

Great, that was going to be my final question.  Sounds like
pseudo-reloc-v2 removes any need to worry.   Since I am planning on
sticking to the released binutils 2.20.1, I'll add that to my LDFLAGS
(to get v2 for 32-bit compilations) and also add -enable-auto-import


Thanks!
Paarvai

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Mingw-w64-public mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public

Reply via email to