❦ 10 août 2015 19:34 +0200, Axel Beckert <abe@debian.org> :
>> > The ieee-data package doesn't offer those files, only the textual
>> > versions. I don't know if the format of those files are standard.
>
> Which ones? the .idx ones or the textual format ones?
The .idx ones.
>> > I generate them at compile-time from oui.txt and iab.txt.
>
> Can they be generated in maintainer scripts instead?
Yes.
>> I think the good way to do depend on ieee-data a register a
>> dpkg-trigger regenerating each time ieee-data change.
>
> Yep, that'd be the best way from my point of view.
>
>> Please change it using a mowe in order to change the file atomically.
>
> The main reason for this tag is to have _one_ single source of this
> data in Debian. We currently have about a dozen of it and some of it
> are really ancient. This also affects the same in data different
> formats than the original one.
>
> arpwatch e.g. seems to use also its own format (which is the reason I
> haven't fixed it there despite I was the last QA uploader of it).
Well, half of the time I spend on packaging of python-netaddr is on
those IEEE files. They change their locations, they get new lintian
_errors_, they are tiresome. Why is it an error and not just an
information?
If I start generating those files at install time, I am pretty sure that
I will get bug because I am generating files outside of /var, because I
don't purge them appropriately, because they are not regenerated when
ieee-data is updated or stuff like that. Building them at compile time
is a fine alternative. The package gets updated about one time per year
and the data is refreshed at this moment.
--
Use the "telephone test" for readability.
- The Elements of Programming Style (Kernighan & Plauger)
Attachment:
signature.asc
Description: PGP signature