Re: Bug#87159: explanation of Build-Depends et. al. is unclear
On Thu, Feb 22, 2001 at 06:15:19PM -0700, Jason Gunthorpe wrote:
>
> On Thu, 22 Feb 2001, Sean 'Shaleh' Perry wrote:
>
> > Specifically, [!i386 m68k] seems like it could be valid, but seems to not be.
> > The archs are also whitespace separated, some people are using commas. Perhaps
> > an exmple with multiple arches would be good.
>
> It is supported by APT..
>
> Python 1.5.2 (#0, Apr 5 2000, 04:40:10) [GCC 2.95.2 20000313 (Debian GNU/Linux)] on linux2
> >>> import apt_pkg
> >>> apt_pkg.init()
> >>> apt_pkg.Config["APT::Architecture"] = "m68k"
> >>> apt_pkg.ParseSrcDepends("apt (> 1.0) [!i386 m68k]");
> [[('apt', '1.0', '>=')]]
> >>> apt_pkg.Config["APT::Architecture"] = "i386"
> >>> apt_pkg.ParseSrcDepends("apt (> 1.0) [!i386 m68k]");
> []
> >>> apt_pkg.Config["APT::Architecture"] = "sparc"
> >>> apt_pkg.ParseSrcDepends("apt (> 1.0) [!i386 m68k]");
> []
APT may parse it, but the example here makes no sense, which is why it
was discouraged. "!i386" means everything but i386, and "m68k" means
m68k. So this example would be equivalent to "[m68k]". IOW, there is no
reason to complicate it by allowing both types to be in the grouping.
FYI, sbuild causes a failure in this example.
--
-----------=======-=-======-=========-----------=====------------=-=------
/ Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \
` bcollins@debian.org -- bcollins@openldap.org -- bcollins@linux.com '
`---=========------=======-------------=-=-----=-===-======-------=--=---'
Reply to: