[tur-users] Please put arch= tag after pkgdesc=

[Robert Emil Berge]

On Tue, 24 Apr 2007 16:06:56 -0700
Jason Chu <jason at archlinux.org> wrote:

> On Tue, Apr 24, 2007 at 11:23:13PM +0200, Robert Emil Berge wrote:
> > On Tue, 24 Apr 2007 16:56:06 +0100
> > "Phil Dillon-Thiselton" <dibblethewrecker at gmail.com> wrote:
> > 
> > > On 24/04/07, Roman Kyrylych <roman.kyrylych at gmail.com> wrote:
> > > >
> > > > I think options should be placed after all other variables, e.g.
> > > > after md5sums. Also please do not place md5sums array at the
> > > > end of PKGBUILDs.
> > > 
> > > Why?  Mate, I got much better things to do than move one line that
> > > hardly anyone will ever look at!
> > 
> > I agree. I suppose the x86_64 people have some reason for insisting
> > on a standard for the arch= tag, but I see no reason for having
> > strict rules for the order of all tags. I place the md5sums at the
> > end of the PKGBUILD because that's where it ends up when I do a e.g 
> > md5sum pkgname*gz >>PKGBUILD
> 
> Actually, this one has hidden errors.  I can't remember the
> specifics, but at one point PKGBUILD was sourced and then a return
> code was checked.  If the build function had an error in it but there
> were md5sums at the end of the PKGBUILD it could cause things to
> break later.  Andy ran into it when doing x86_64 building a lot time
> ago.
It could be something that has happened to me quite a lot with Lowarch.
I do the md5sum thing, which gives the sum at the end of the PKGBUILD
and then the file name. Sometimes I forget to remove the file name at
the end of the file. The packages builds ok, but when doing gensync,
the process will break, because you get an error at the end of the
PKGBUILD. The solution is simply not to forget to remove the filename
after the md5sum.