[Rubygems-developers] Another plug for Simon's patch

[Mauricio Fernández]

On Thu, Apr 01, 2004 at 07:28:01AM -0500, Jim Weirich wrote:
> I'm not against a versioning policy.  However, as I tend to have an 
> EnablingAttitude (http://martinfowler.com/bliki/EnablingAttitude) rather 
> than a DirectingAttitude 
> (http://martinfowler.com/bliki/EnablingAttitude), I would phrase the 
> policy as a suggestion and then point out the benefits of following the 
> policy (e.g. if you manage your versions in /this/ way, then users can 
> reliably use the version comparison tools in RubyGems to manage what 
> versions are used).  At that point I would let the individual developers 
> make their own choices.

I would *strongly recommend them* to follow the policy, and to break it only if
 * they know the consequences
 * they have a good reason for it

> Now, exactly /what/ policy to adopt is another question.  I've read the 
> example for library-0.0.0, but it's not clicking for me.  I'm not sure 
> why, let me think on it a bit and I'll respond later.
>

I also dislike the fact that it's much 'heavier' than the other
minimalistic policy I talked about. But it's really nice in the sense
that it allows you to have an arbitrary number of
"backwards-compatibility levels", and that causally (the gemmaker need
not know which future releases will be compatible).

> There is a related issue here.  I believe the current behavior is to 
> automatically download any missing dependencies.  I do have a small 
> problem with the entirely automatic download.  I really would like to be 
> notified of the gems (and version) that are needed and would like to 
> explicitly approve them.  Debians apt-get does something like that. 
> When I install a package, it lists the dependent packages it needs to 
> download and asks for a go/no-go decision.  I think that would be a good 

I believe that still puts too much burden on the end-user; he will have
to know which versions are incompatible with the stuff he's got
installed. He might break things and fail to realize it until *much
later*, at a time where he might be unable to relate that to the
installation that caused it all (esp. if several other were performed
in between).


-- 
Running Debian GNU/Linux Sid (unstable)
batsman dot geo at yahoo dot com

'Ooohh.. "FreeBSD is faster over loopback, when compared to Linux
over the wire". Film at 11.'
	-- Linus Torvalds