libtool-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Minor change for cygwin, libtool-1.5.22


From: Ralf Wildenhues
Subject: Re: Minor change for cygwin, libtool-1.5.22
Date: Mon, 24 Apr 2006 09:56:28 +0200
User-agent: Mutt/1.5.11

Hi Charles,

* Charles Wilson wrote on Sun, Apr 23, 2006 at 11:31:35PM CEST:
> I ran across an issue when working with CVS gettext on cygwin.  When 
> building objects destined for inclusion in a DLL (so-called PIC objects, 
> even tho all .o's are "position independent" on MSWin), cygwin ought to 
> define DLL_EXPORT just like mingw.  Even though we don't really need to 
> for auto-import, there are projects that try to do things the Microsoft 
> Way with declspec, instead -- and use --disable-auto-import.  As it 
> happens, gettext post-0.14.5 will be one of those.

Ah, yes, there was some earlier message of you floating around somewhere:
http://lists.gnu.org/archive/html/bug-gnu-utils/2006-04/msg00022.html

Patch looks good; interestingly:

> @@ -5185,9 +5189,11 @@
>         esac
>         ;;
>  
> -    mingw* | pw32* | os2*)
> +    mingw* | cygwin* | pw32* | os2*)
>        # This hack is so that the source file can tell whether it is being
>        # built for inclusion in a dll (and should export symbols for example).
> +      # Even tho PIC is default for cygwin, still need this for old-style
> +      # (--disable-auto-import) libraries
>        _LT_AC_TAGVAR(lt_prog_compiler_pic, $1)='-DDLL_EXPORT'
>        ;;
>  

This last bit is for non-gcc compilers; so that should've been there
before in fact.  So far Libtool has sometimes assumed that $GCC != yes
means we're using MSVC, and thus LINK for linking (thinking of Peter's
pending patches for MSVC support); but we need to move away from this
line of thought if we want to support PGI on w32 (IIRC there was a
request to this end a while ago).

I'll apply this later (and let you know), probably without the comment
added to that last hunk.

Cheers, and thanks,
Ralf




reply via email to

[Prev in Thread] Current Thread [Next in Thread]