[csw-pkgsubmissions] newpkgs ruby19, ruby19dev, ruby19ri, ruby19sa(...)

Philip Brown phil at bolthole.com
Mon Apr 19 21:25:24 CEST 2010


On Wed, Apr 14, 2010 at 4:56 PM, Ben Walton <bwalton at opencsw.org> wrote:
> Excerpts from Ben Walton's message of Thu Apr 08 21:35:53 -0400 2010:
>
> Bump.
>
>> * ruby19: new package
>>   + ruby19-1.9.1p376,REV=2010.04.03-SunOS5.9-i386-CSW.pkg.gz
>>   + ruby19-1.9.1p376,REV=2010.04.03-SunOS5.9-sparc-CSW.pkg.gz
>>   + ruby19dev-1.9.1p376,REV=2010.04.03-SunOS5.9-i386-CSW.pkg.gz
>>   + ruby19dev-1.9.1p376,REV=2010.04.03-SunOS5.9-sparc-CSW.pkg.gz
>>   + ruby19ri-1.9.1p376,REV=2010.04.03-SunOS5.9-all-CSW.pkg.gz
>>   + ruby19samples-1.9.1p376,REV=2010.04.03-SunOS5.9-all-CSW.pkg.gz
>>   + ruby19tk-1.9.1p376,REV=2010.04.03-SunOS5.9-i386-CSW.pkg.gz
>>   + ruby19tk-1.9.1p376,REV=2010.04.03-SunOS5.9-sparc-CSW.pkg.gz
>

sigh.

Sorry I didnt reply about this sooner.

i HAVE been thinking about this.
(Deep, dark, unhappy thoughts, about how more "open source" developers
badly need to gain a clue about maintainability, sustainability, and
business/stability-friendly practices....)

I'm thinking that for ruby, what sounds best for us, would be to
migrate to something similar to our berkeleydb practice.
Which is approximately as follows:

- we maintain version-specific packages of ruby
- we maintain them only so long as WE need them for OUR packages
- once none of OUR packages uses an older version, we drop it
- there is no generic "ruby" package, since it is guaranteed to cause
problems in the future due to shortsightedness of the ruby developers.

So... we COULD replace all our "CSWruby*" packages, with "empty"
packages that depend on ruby18 packages(that you would then have to
make)... but only as a temporary stopgap measure. The packages need to
clearly identify that they are a transition package only.
There will be no ongoing "CSWruby" package, after transition is complete.

Sound okay?


More information about the pkgsubmissions mailing list