Re: config.sub, upstream author responsibility?

[Raja R Harinath <harinath@cs.umn.edu>]

Hi,

Adrian Bunk <bunk@fs.tum.de> writes:
> On Fri, 1 Jun 2001, Alwyn Schoeman wrote:
>> I just got a bug report on one of my packages that it doesn't recognise a
>> specific hardware platform and requires a new config.sub.  As this
>> is part of the original source code, is this the upstream author's
>> responsibility? 
>>
>> If so, is the only thing I can do to request an update by that author or can
>> I actually go change it?
> 
> You should request an update by the upstream author, but (at least for
> DFSG programs) you should update the file yourself when there's no
> upstream version with the updated file in the near future - replacing
> config.sub and config.guess with newer version is usually a save change to
> a package.

But not always.  For example, there was a time last year that
'config.guess' used to identify x86 GNU/Linux machines as, say

   i686-unknown-linux

while it currently identifies them as

   i686-unknown-linux-gnu

There are lots of 'configure.in's that have

  case $host in
  *-linux-gnu) ...;;
  esac

So, casual upgrades of config.guess are usually safe, but not always.
'config.sub' maybe safer to upgrade, though.

- Hari
-- 
Raja R Harinath ------------------------------ harinath@cs.umn.edu
"When all else fails, read the instructions."      -- Cahn's Axiom
"Our policy is, when in doubt, do the right thing."   -- Roy L Ash