As motu-release has decided to revert the upon, I'm marking this fix
committed. That being said, we still need to find a sane alternative, so
I don't believe we should close this bug until one exists.

Looking at the other interpreters in Debian and Ubuntu, we don't support
the usage of CPAN or PEAR since it installs things without the package
systems knowledge, nor do they get properly autoremoved if you remove
the interpreter itself. Neither CPAN or PEAR has been modified to
interact with APT, which would be a very complicated change, and the
wrong one to make.

In this case, I believe the correct course of action is to depreciate
the gems package manager (perhaps modify it  to display a giant
warning), but leave it available so a user can use it if they choose to
override our choices, and then package gems directly into APT. I'm
assuming gem installation from source packages are relatively
striaghtforward (something like perl Makefile.PL; make; make install, or
python setup.py), so it is likely possible to create a template set of
files for debian/, allowing for rapiding packaging of every gem.

Now looking at Rubyforge, there are roughly 7,000 gems, so packaging
each one individually may be a little close to impossible. Still, the
most common, and most popular ones should be packaged and archived;
Rails should be in the archive for instance.

No matter what we will do, people will still have the opportunity to
shoot themselves in the foot by installing the source packages or using
gems over APT. However, by providing a sane alternative, we can avoid
that possibility.

** Changed in: libgems-ruby (Ubuntu)
       Status: Confirmed => Fix Committed

-- 
rubygems bin in PATH potentially breaks other applications and violates all 
sense of decency in packaging.
https://bugs.launchpad.net/bugs/262063
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to