[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#168761: Documentation is quite messy.



I really don't know why i insisted with debconf-devel(8) while it is in
section 7, but...

On Tue, Nov 12, 2002 at 02:12:33PM -0500, Joey Hess wrote:
> This is because the text data-type is kind of unused and ill-supported
> and deprecated. I have added a mention of it anyway.
> 
> The debconf specification is, as its title might suggest, authoratative.
> The debconf-devel(7) man page is descriptive. And yes, it does refer
> readers to the debconf specification for definitive docs, in at least 3
> places.

It's kindly intuitive since is among the policy documentation, but the man
page is distributed with the debconf source that makes it enough
authoritative: confused by the differences i've found and i wandered which of
these documents where up to date.

> This is mildly deprecated, and will be replaced with a better mechanism
> in the future, if I can ever get anything through the policy process.

I'd like to know something about this: i want to use variable in
Choices field, but i'd like to save some time.

> I recently removed most of the details from debconf-devel(7) (which does
> mention -ll fields still), since the internals of how the fields are
> named is all handed by po-debconf now, like the man page says.

Don't you think it's a good idea to document this anyway?
BTW I think that the one of this documents (i'd say the specifications) should
document the list of availables flags.

> debconf-devel(7) is there to be a single source that a package
> maintainer should be able to read to learn pretty much everything they
> need to know to use debconf. As such its descriptions of the debconf
> protocol and templates file and config script are very important. Each
> of these sections contains a lot of information and answers to FAQs that
> will never get into the policy document. I will not be removing them. I
> don't expect that keeping the few bits that overlap with the spec in
> sync will be a big deal, since I have to update debconf's code anyway if
> the spec changes.

I understand your point, so this bug may be considered as a request to sync
the debconf documentation as soon as possible.

> I disagree. It's much nicer to have a single page that I can refer
> developers to, and which can be read right though or used as a reference
> than it is to have the docs scattered amoung many pages. I have gotten
> many fewer repeated FAQs from developers about debconf since I wrote the
> single large man page.

It's nicer to me too, but here there is the problem of keeping two distinct
souce in sync and they seems not to have the same frequency. With
this sentence i'm not blaming anyone for this. My offer still stands, even if
i suppose i'd contact Randolph Chung first.

ciao,
-- 
Luca - De Whiskey's - De Vitis              | Elegant or ugly code as well
aliases: Luca ^De [A-Z][A-Za-z\-]*[iy]'\?s$ | as fine or rude sentences have
Luca, a wannabe ``Good guy''.               | something in common: they
local LANG="it_IT@euro"                     | don't depend on the language.

Attachment: pgpzUaclgoIPO.pgp
Description: PGP signature


Reply to: