JRuby

Provide easy mechanism to uninstall RubyGems

Details

  • Type: Improvement Improvement
  • Status: Closed Closed
  • Priority: Major Major
  • Resolution: Won't Fix
  • Affects Version/s: None
  • Fix Version/s: None
  • Component/s: Extensions
  • Labels:
    None
  • Number of attachments :
    0

Description

Recent flurry of RubyGems updates have left bitter taste in some users' mouths.

Consider removing RubyGems from our distribution, so that we can work with alternative(s) such as SlimGems (https://github.com/slimgems/slimgems).

Activity

Ascending order - Click to sort in descending order
Hiro Asari made changes -
Field Original Value New Value
Component/s Extensions [ 12281 ]
Component/s Core Classes/Modules [ 12230 ]
Hide
Permalink
David Calavera added a comment -

+1

Show
David Calavera added a comment - +1
Hide
Permalink
Thomas E Enebo added a comment -

So the idea is to make users download a gem soln 100% of the time versus installing something which works for some percentage (majority?) of people? I would like to support installing gems in some fashion out of the box without requiring an additional download.

Could we possibly find a better way of allowing people to override our choice so that if you want slimgems you can at least remove rubygems?

Show
Thomas E Enebo added a comment - So the idea is to make users download a gem soln 100% of the time versus installing something which works for some percentage (majority?) of people? I would like to support installing gems in some fashion out of the box without requiring an additional download. Could we possibly find a better way of allowing people to override our choice so that if you want slimgems you can at least remove rubygems?
Hide
Permalink
Hiro Asari added a comment -

Yes, I am OK with an easy RubyGems uninstall mechanism. I'm editing the summary now.

Show
Hiro Asari added a comment - Yes, I am OK with an easy RubyGems uninstall mechanism. I'm editing the summary now.
Hiro Asari made changes -
Summary Drop RubyGems Provide easy mechanism to uninstall RubyGems
Hide
Permalink
Yoko Harada added a comment -

Ruby 1.9 is shipped with RubyGems. If the change doesn't affect any compatibility to CRuby version, it is good to substitute to slimgems or other. No harm?

Show
Yoko Harada added a comment - Ruby 1.9 is shipped with RubyGems. If the change doesn't affect any compatibility to CRuby version, it is good to substitute to slimgems or other. No harm?
Hide
Permalink
Hiro Asari added a comment -

It must be noted that maven integration should still work whatever solution we find on this issue.

Show
Hiro Asari added a comment - It must be noted that maven integration should still work whatever solution we find on this issue.
Hide
Permalink
Thomas E Enebo added a comment -

Hiro

Yeah, SlimGems may work for now with our RG maven patching since it is just a fork, but over time I am worried. What we really need is for RubyGems to accept some patches to allow us to hook Maven support in. Right now is probably a bad time to bring it up to them.

To date the issues with allowing uninstalling (or just plain removing) RG is:
1. 1.9 ships RG and we support 1.9 compatibility
2. Maven support is monkey-patching RG to work. If you uninstall and replace then this feature is very likely to just be broken.

Anything else? It seems this is getting dangerously close to WONTFIX (at least until the RG landscape changes to re-evaluate the idea).

Show
Thomas E Enebo added a comment - Hiro Yeah, SlimGems may work for now with our RG maven patching since it is just a fork, but over time I am worried. What we really need is for RubyGems to accept some patches to allow us to hook Maven support in. Right now is probably a bad time to bring it up to them. To date the issues with allowing uninstalling (or just plain removing) RG is: 1. 1.9 ships RG and we support 1.9 compatibility 2. Maven support is monkey-patching RG to work. If you uninstall and replace then this feature is very likely to just be broken. Anything else? It seems this is getting dangerously close to WONTFIX (at least until the RG landscape changes to re-evaluate the idea).
Hide
Permalink
Hiro Asari added a comment -

The more I think about it, the more danger I see in pursuing it at this point, unfortunately.

Maybe the situations will improve with RubyGems and it becomes feasible in the future.

I'll close this for now, and revisit in the future.

Show
Hiro Asari added a comment - The more I think about it, the more danger I see in pursuing it at this point, unfortunately. Maybe the situations will improve with RubyGems and it becomes feasible in the future. I'll close this for now, and revisit in the future.
Hiro Asari made changes -
Status Open [ 1 ] Resolved [ 5 ]
Resolution Won't Fix [ 2 ]
Charles Oliver Nutter made changes -
Status Resolved [ 5 ] Closed [ 6 ]

People

  • Assignee:
    Hiro Asari
    Reporter:
    Hiro Asari
Vote (1)
Watch (3)

Dates

  • Created:
    Updated:
    Resolved: