so what is the consensus?
we have people who care and wish there was versioning (in whatever most practical way this is realized)
some who say which is the only possible way to do it and immediately say they will not comply
some how all their negativity aside agree it would be lack of compliance that would hinder the success of the idea
so can this be summed up that we as gentoo developers do not care enough to do something that would allow us to prevent major breakage for users?
that we do not care enough to demonstrate enough professionalism "in the wild" to get people ever interested into something corporations (whatever the final name of the product would be) would be interested in testing?
that we dont care enough that we have a process that costs fellow developers money when a demonstration of gentoo simply doesnt work, and the interested customer in turn be turned off by what they see?
is this how we want to present ourselves?
someone please tell me this is not what this thread shows, cause when i go through it all, this is what i read.
i think this thread should continue under another name, and i will make sure the new topic will be shorter and more to the point, another, new point, just related, and i will try to no longer reply to pointless, purely negative, contructive criticism free comments, maybe those highly annoyed devs could simply refrain from creating traffic unless they have something to add, many should appreciate that, most likely..quietly
Properly signed and key available
Sincerly,
Daniel Goller
Ciaran McCreesh wrote:
On Tue, 18 Jan 2005 15:11:14 +0000 Stuart Herbert <[EMAIL PROTECTED]>
wrote:
| On Tuesday 18 January 2005 08:28, Ciaran McCreesh wrote:
| > Version specific code is already doable with versionator. | | I maintain a couple of eclasses, and personally I find it easier to
| maintain separate eclasses for the different releases where things
| substantially change. So I don't version eclasses in the way that
| Daniel suggests, but neither do I subscribe to Ciaran's
| one-size-fits-all-really-it-does-and-all-
| other-solutions-are-wrong-because-I-say-so view of the world (I assume
| it's his exam stress making him worse than normal).
No, it's the increase in nonsense on this list that's making me worse than normal.
| It also has the benefit that the older ebuilds/eclasses don't get
| changed just to accomodate new (and incompatible) changes. Makes
| things more stable for everyone.
| | I'm sure versionator has its uses in ebuilds (currently used by 24
| packages), but I don't find it a sound engineering solution to all the
| ills of eclasses.
Sure, if you're going to make huge changes for different major versions of the package, have a foo-1 and a foo-2 eclass. This doesn't involve any kind of portage versioning, however, and so is totally irrelevant to the discussion.
signature.asc
Description: OpenPGP digital signature
