| Summary: | Please unmask ruby-1.8.7 | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Hanno Böck <hanno> |
| Component: | [OLD] Development | Assignee: | Gentoo Ruby Team <ruby> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | caleb |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 260760 | ||
|
Description
Hanno Böck
2008-12-23 09:50:55 UTC
Ruby 1.8.7 has a number of incompatibilities compared to ruby 1.8.6, which means that the software is not a seamless upgrade from 1.8.6. At the same time it is also not possible to install ruby 1.8.6 and ruby 1.8.7 in parallel. The consensus amongst distributions seems to be to stick with ruby 1.8.6 at least for now so that existing software won't break. I've updated the package.mask description accordingly. So I think we'll keep ruby 1.8.7 masked for now. I've also cc'ed Caleb who added the ebuilds in case he has more insight. I don't have a compelling argument one way or the other. I think keeping it masked now is okay, if we need to update the package mask with a more compelling reason then I'm happy to do that. At least debian and opensuse already use ruby 1.8.7. I'd suggest for broken software to create a tracker bug (like it's commonly done for all big updates that tend to break lots of stuff) and open bugs for software not running in ruby 1.8.7. That'd give some insight what's blocking the unmasking and give devs a hint where they could help to speed up things. I think one problem here is that because of rubygems we cannot rely on FEATURES=test to identify the broken packages, which makes it much harder to identify the issues. On the other hand, if we decide that 1.8.7 test is a goal for us, I can do my best to test packages in my tinderbox, manually where needed. Given that we have had to fix a number of problems with ruby 1.8.6 ourselves which got fixed upstream only in 1.8.7, and given that other distributions are also moving to 1.8.7, I think we should also start to think about unmasking ruby 1.8.7. In fact, I've just installed it on my test machine to see what really breaks. I've added a revision which uses eselect ruby rather than ruby-config (since I wanted to test it too, and my system is already converted to eselect ruby). Okay, the mask is gone. Let's see what happens ;) Closing this bug since ruby 1.8.7 is now unmasked. |
