On Mon, Nov 25, 2002 at 03:25:27PM +0000, Tim Bradshaw wrote:
> Has anyone succeeded in getting this to build?
>
> I get it to configure OK, but then it dies trying to make dependencies
> (I think) in common-src/debug.c, with errors that look like pid_t and
> various other important types are not known (and also, possibly,
> prototypes are missing / don't match definitions, which may be because
> the P macro isn't defined.
>
> Everything looks OK - it thinks it should include unistd.h (which is
> where pid_t is defined), but if I trace what gcc is actually including
> it seems to include (the fixincluded) unistd.h but *not* the files it
> includes, (even though _UNISTD_H is not defined...), which smells
> like some gcc bug. But I can't work out what it is, and it seems
> ludicrously unlikely that gcc has such an obvious bug.
>
> I have 2.4.2p2, but built with an older gcc (probably 2.9x or maybe an
> early 3, certainly not 3.2). I've just rebuilt it with 3.2 and it
> works fine (well, it builds with no complaints, I don't want to
> install it).
>
> All I can think of now is to gradually whittle away at amanda.h /
> config.h or compare them with the 2.4.2p2 ones until something
> changes, but I thought I should ask first (or, realistically just
> stick with 2.4.2p2...).
>
> The configure line I'm using is:
>
> ./configure --prefix=/local \
> --with-user=root --with-group=backup \
> --with-config=tfeb-backup \
> --with-gnutar=/local/gnu/bin/tar \
> --with-libraries=/local/gnu/lib \
> --without-debugging
>
> which should be pretty unexceptionable I think.
--without-debugging is unusual.
Try this patch.
Jean-Louis
--
Jean-Louis Martineau email: martineau AT IRO.UMontreal DOT CA
Departement IRO, Universite de Montreal
C.P. 6128, Succ. CENTRE-VILLE Tel: (514) 343-6111 ext. 3529
Montreal, Canada, H3C 3J7 Fax: (514) 343-5834
amanda.without-debugging.diff
Description: Text document
|