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

Re: Please allow the migration of dhelp



Julien Cristau dijo [Mon, Aug 20, 2012 at 09:51:13PM +0200]:
> > > #679300 is about the removal from sid.  Removal from testing won't
> > > happen as long as dhelp still depends on it.
> > 
> > Ok - But, the fact is that in testing you install dhelp, it won't work
> > at all as ruby-commandline fails silently under the default Ruby
> 
> That's only because somebody decided to change the default Ruby version
> after the last minute.
> 
> > version. Thus my earlier request to allow for the newer dhelp to
> > migrate into testing, removing the dependency on the buggy
> > ruby-commandline, and allowing the later to be removed painlessly.
> 
> Seems to me the minimal fix for wheezy is to make sure dhelp /
> ruby-commandline use ruby 1.8.

There is another way out - dhelp is currently at 0.6.20 in testing and
0.6.21+nmu1 in unstable. I wanted to avoid to suggest a +wheezy1
release, just fixing this trivial dependency issue, but am willing
to - I feel it impractical, though.

The solution you mention is not realistic, IMO. If a user has any
Ruby-using stuff installed in his system, he will have Ruby 1.9.1
installed by default. Then, he installs dhelp, which requires 1.8, and
*the whole of* 1.8 is brought in. Dhelp will then need to refer
explicitly to /usr/bin/ruby1.8,

I agree with you, the decision to switch the Ruby release for Sid was
hasty, and I would have prefered for it to wait. But it is a done deal
- And the advantages are many. Ruby 1.8 is no longer supported
upstream. The language's speed has hugely improved. And many newer
modules are no longer compatible with 1.8 (including several packaged
already for Debian). 

So, please consider this again. dhelp 0.6.21 was uploaded three weeks
before the freeze — I don't know why it was not accepted in time. The
bug I refer to (#678055) was closed then — I only did this last upload
on July 2 because, even though dhelp had been patched with the changes
I sent (and had many other improvements clearly aimed at entering
wheezy), but the *real* (i.e. behaviour) bug had been fixed
already. If it is *not* accepted, I (well, much rather, Georgios who
is the maintainer) can just apply the fix for #678055 on 0.6.20 as
0.6.20+wheezy1 — But I don't think that's making a difference for the
good of the users.

All in all, I just want to be able to drop a badly buggy
ruby-commandline which has no sense in Wheezy.

Attachment: signature.asc
Description: Digital signature


Reply to: