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

Re: mass-commit: dependencies

On 12-07-06 at 10:54am, gregor herrmann wrote:
> On Fri, 06 Jul 2012 01:17:47 +0200, Jonas Smedegaard wrote:
> [example: libanyevent-xmpp-perl]
> >  c) change to build-depend versioned on debhelper
> >  c) change to build-depend unversioned on cdbs
> > 
> > c) is almost harmless and needless: Yes, Joey instructs to do so but 
> > really a dependency on debhelper 7 is implicitly satisfies as no 
> > older version is provided in any currently supported release of 
> > Debian. APT dependency resolving is already plenty burdened - we 
> > really should work actively to _reduce_ needless versioned 
> > dependencies.
> So here you are arguing that the version is unnecessary because 
> there's nothing older around ...
> > d) is harmful: a recent cdbs _is_ needed.  In this particular case 
> > the CDBS snippet perl-makemaker.mk is included which was unavailable 
> > prior to cdbs 0.4.73.
> ... and here you point out that the version is necessary, although 
> stable has 0.4.89 (and oldstable is archived [and I can't check it's 
> value right now]).
> Just trying to understand ...

You are quoting a waaaaay old email of mine.  Back then oldstable was 
still relevant (in my head). :-P

> > I just feel here concretely that it is too noisy and invasive with 
> > no (in my tiny test) actual benefit :-/
> The real benefits are e.g. with dual-lived modules, where it nicely 
> cleans several versions of "libfoo | perl >" (or the other way round) 
> cases, turning them around when a newer perl is available or removing 
> them when the oldest perl has the module in core etc.
> And for cleaning up large packages with tons of versioned dependencies 
> where the versions are not needed anymore.

I see the benefits.  And I trust the sanity in the resolving now.

Feel free to apply it also to CDBS-based packages when updated to not 
add unneeded debhelper versioning.

I'll revert stylistic changes when I update control file via control.in.  
I recommend to only push changes when _semantic changes occur, not only 
for stylistic changes.  Not sure how that could be automated - I still 
think Config::Model ought to handle that.

> Cheers,
> gregor, on the nice balcony but without wifi


 - Jonas

...on the couch in the new flat on the island, wifi not yet plugged in

 * 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: