Re: practices for porters (was Re: large number of source package ...)
On Sun, Sep 27, 1998 at 04:40:48PM -0400, Adam P. Harris wrote:
> Paul Slootman <paul@wau.mis.ah.nl> writes:
> > The way I do uploads for Alpha where patches were necessary, is to only
> > upload the binary, and to submit a patch to the BTS. This way
> > (a) the package maintainer gets a chance to integrate the patch in the
> > way he sees fit (e.g. maybe he can see a cleaner way of fixing something),
> > (b) the source for the NMU binary can be reconstructed from the standard
> > sources plus the patch from the bug report.
> > I find this works very well. The only drawback is where the maintainer
> > weeks if not months in integrating the patch.
>
> Does this practice receive the official blessing of the archive
> maintainers?
It has one definite minus - if Paul does that, and then I discover it
also needs a powerpc patch, we stumble on top of each other more likely
than not.
But it has one definite plus - it's more effective than my method. I
generally file a bug with my patches, and wait. I'm just about at the
point in my cycle where I do NMUs of the packages currently in my
queue, as soon as I find the time.
> A section addressed to i386 people giving suggestions to maximize
> liklihood of an easy port would be nice. I have learned some things
> slowly and now have to go back and check all my packages.
> eg, The binary named 'egcc' is not available for all
> architectures.
Actually, it is if egcs is installed at all, in the recent versions.
And if egcc is specified, then the presumption is that the most recent
version is probably a good idea.
Dan
Reply to: