[csw-maintainers] New catalog "unstrict"

Dagobert Michelsen dam at opencsw.org
Sun Dec 14 22:23:47 CET 2008


Hi James,

Am 14.12.2008 um 11:00 schrieb James Lee:
> On 13/12/08, 12:54:16, Dagobert Michelsen <dam at opencsw.org> wrote  
> regarding
> [csw-maintainers] New catalog "unstrict":
>
>> I want to pick up the thread on a more open catalog
>> ("unstrict/" or "open/") which is a superset of current/
>> where it is not necessary to provide versions for all
>> Solaris releases or sparc/x86. Being able to migrate
>> freely between sparc and x86 and being able to upgrade
>> Solaris versions for current/ I consider very important.
>> The separate catalog should make clear that this strict
>> policy is not valid when using the other catalog. That
>> means "unstrict" contains all packages from current, but
>> additional ones like cvsup, opera, tme etc.
>
> "unstrict" isn't a word, I guess "lenient" is the closest although
> perhaps "noncompliant" is what you mean.

Someone said this term is used at some Linux distribution.
Trygve? William? Sorry, can't remember.

> It's a slippery slope.  The deal is that to share in the efforts of
> others one has to make the effort to support others perhaps in areas
> that are of no direct need to yourself.  The indirect need is that
> your contribution become part of a wider entity that adds weight to
> your own contribution and the whole.
>
> Where there is a case for a particular package we should allow
> asymmetry.  However there has to be a valid case, not just "it didn't
> compile first time". Example could be a tool to analyse Solaris
> 10 only features (zones, ZFS, dtrace) but take care that the depends
> don't knock out otherwise useful packages.

We already have dtracetoolkit. How about the vendor does not make
a version (like Opera Solaris 8 x86, right?) or there is no
assembly backend (like for CVsup Solaris x86). Theoretically this
could be written, but is a much bigger work than tweaking flags
or writing a couple-of-lines patch.


Best regards

   -- Dago



More information about the maintainers mailing list