Bug#142164: Packages files should be in UTF-8

[Goswin von Brederlow <brederlo@informatik.uni-tuebingen.de>]

Kurt Roeckx <kurt@roeckx.be> writes:

> On Thu, Sep 29, 2005 at 11:32:13PM -0500, Manoj Srivastava wrote:
>> reassign 142164 general
>> thanks
>> 
>> Hi,
>> 
>>         Just because a topic has been discussed on the policy
>>  discussion list is not reason enough to assign the bug to
>>  policy. 
>
> Note that bug has been reassigned to the policy in 2003.  Spam
> closed it, and it was just reopened.
>
>>  This has nothing to do with creating a package, and certainly
>>  not the place of policy  to lead out by mandating stuff. Again, this
>>  problem needs to be worked out abd put into effect by the bits that
>>  put the Package ile together, and once we have a working solution, we
>>  can start examing the bits of policy that may need to be changed for
>>  the solution.
>
> I agree that the policy shouldn't mandate that the Packages
> file should be encoded in UTF-8.  But I think it should say
> the the control and changelog files are, and I believe that is in
> the scope of the policy.  This would of course have as side
> effect that the Packages file ends up as UTF-8 too.
>
> Note it says that is is "recommended" to encode the changelog in
> UTF-8 (C.2.2), it does not say any such thing about the control
> file.  I think we should say that both should be encoded in
> UTF-8, or maybe a must.  But we can change that to a must at a
> later date.
>
>
> Kurt

I think all six, dsc, changes, changelog, control, Packages and
Sources files, must be UTF-8 for three simple reasons:

- They all interconnect through tools that preserve / ignore the
  encoding.
- They can't have different encodings since guessing the right
  conversions would be hell.
- UTF-8 is the only practical format covering all cases.

If policy has something to say about one of them then it has to say
something about all of them.

MfG
        Goswin