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

Re: mass-commit: dependencies



On 12-07-07 at 05:08pm, Dominique Dumont wrote:
> Le Saturday 7 July 2012 01:07:44, gregor herrmann a écrit :
> > > Personally, I would recommend keeping [debhelper versioned 
> > > dependency], but IIRC Jonas and I have had this argument before.  
> > > :)
> > 
> > hihi
> > 
> > i also prefer to keep it but maybe C::M could at least not add it if 
> > it's not there but keep it if it's there?
> 
> ok. that's what I'll implement except when debian/compat is '9'. In 
> this case, cme will "fix" debhelper dep so as to have debhelper >= 9
> 
> Is this fine ?

Yes.

I assume you don't mean hard-coding to special-case debhelper 9 in 
particular, but b) below...


I believe these are fine:

 a) Tighten versioning to match compatibility level.
 b) Introduce versioning to match compatibility level, when needed 
    version is missing from a currently supported Debian distribution 
    release
 c) Drop versioning, when needed version is satisfied in all currently
    supported Debian distribution releases, and Debian Policy does not
    dictates differently, and packaging documentation does not request
    differently

I believe these are not fine, as we favor different conflicting ones 
(but would be cool to have available as optional flags):

 d) Introduce versioning (e.g. to match compatibility level), when 
    needed version is satisfied in all currently supported Debian 
    distribution releases
 e) Drop versioning, when needed version is satisfied in all currently
    supported Debian distribution releases, and Debian Policy does not
    dictate differently (disregarding package documentation requests)

These are IMO bad, as they can cause breakage in some situations (but 
would be cool to have as warnings and/or optional flags):

 x) Bump compat level to match versioning
 y) Relax versioning to match compatibility level


Beware that Debian in general only care about FTBFS (Failure To 
Build-against-same-distro-release-or-unstable From Source), so even if 
the Perl team agrees to also care about FTBBFS (Failure To 
Build/Backport From Source) it might be wise for Config::Model to allow 
suppressing FTBBFS detection or offer it only optionally (which the Perl 
team would then add to its group-wide scripts.


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private

Attachment: signature.asc
Description: Digital signature


Reply to: