Re: pkg/42102 - Adjusting __weak_alias in libnbcompat

[David Brownlee <abs%netbsd.org@localhost>]

Joerg - would you be happy with the '#ifdef __weak_alias'  being '#if
defined(__NetBSD__) && defined(__weak_alias)', or do we need to handle
more OS variants?


2009/12/31 OBATA Akio <obache%netbsd.org@localhost>:
> The following reply was made to PR pkg/42102; it has been noted by GNATS.
>
> From: "OBATA Akio" <obache%netbsd.org@localhost>
> To: gnats-bugs%netbsd.org@localhost
> Cc:
> Subject: Re: pkg/42102 (pkgsrc fails bootstraping on openbsd amd64)
> Date: Thu, 31 Dec 2009 11:39:05 +0900
>
> ÂOn Thu, 31 Dec 2009 05:35:11 +0900, <abs%netbsd.org@localhost> wrote:
>
> Â> I wonder if this is related to the #ifdef __weak_alias section or an 
> existing
> Â> definition of DB...
> Â> As a quick test could you try editing 
> pkgsrc/pkgtools/libnbcompat/files/db/db/db.c
> Â> and remove the following section around line 48
> Â>
> Â> #ifdef __weak_alias
> Â> __weak_alias(dbopen,_dbopen)
> Â> #endif
> Â>
> Â> then try rerunning bootstrap?
>
> Â__weak_alias also in db/mpool/mpool.c.
>
> ÂAs far as I know, I came from a bug of __weak_alias on OpenBSD
> Â(or you could say that the spec of __weak_alias on OpenBSD is
> Âdiffer than one on NetBSD).
>
> ÂI think that macros prefixed with "__" are not portable, should not be used 
> in
> Âapplications/libraries, or at least, __NetBSD__ macro should be checked 
> first.
>
>