Right now the Patch Viewer supports only CVS, and CVS isn't nearly as popular as it used to be. We should use VCI (a Perl module that I wrote, it's on CPAN) to allow the Patch Viewer to interact with multiple VCSes. VCI would just need a few more features, and then it could possibly entirely replace PatchReader in our code.

I'm all for that. PatchReader's been unmaintained for a couple years anyway.

(In reply to comment #0) > We should use VCI (a Perl module that I wrote, it's on CPAN) to allow the Patch > Viewer to interact with multiple VCSes. It would be fine if VCI didn't have tons of dependencies. I tried to install it once, and gave up pretty quickly.

(In reply to comment #4) > It would be fine if VCI didn't have tons of dependencies. I tried to install it > once, and gave up pretty quickly. I used install-module.pl to install it fine just the other day. It may help to update your CPAN module.

I see no movement on this bug: Is there a viable alternative to VCI?

(In reply to comment #7) > I see no movement on this bug: Is there a viable alternative to VCI? It's not an issue of whether there's an alternative to VCI--VCI is what we would use for this, for sure. The issue is that the work simply hasn't been done yet. Very possibly the Patch Viewer itself will be entirely superseeded by Splinter, the new patch review extension, and then this will be a feature of Splinter instead.

What is some further-reading on Splinter, it's features and when it'll be merged?

(In reply to comment #9) > What is some further-reading on Splinter, its features and when it'll be > merged? bug 570786

That bug is about getting Splinter into B.G.O, not getting it (and more specifically version control support) into Bugzilla (not just bugzilla.mozilla.org) - Bugzilla is used by many organizations/websites: See http://www.bugzilla.org/installation-list/ So basically why are we waiting on Splinter (especially as I've seen no mention of VCS on any documentation I've read about it) if it won't be standard to Bugzilla?

We're not waiting on anything, but once the extension is publicly available (which will happen after it is on bmo) then perhaps many people will get this feature as a part of Splinter in the future, and it will be less important for us to implement it. As you can see, this bug was filed in 2007, but we have no worked on it yet. It's not because we're waiting for something--it's that there aren't an unlimited number of programmers available who are working on Bugzilla. If you want to implement the feature described in this bug, we would absolutely welcome your contribution. Our development process is described here: http://wiki.mozilla.org/Bugzilla:Developers

Would this be simply adding the extension and the library to the bzr branch?

(In reply to comment #13) > Would this be simply adding the extension and the library to the bzr branch? No, it would involve a tremendous amount of code work. VCI is not a Bugzilla Extension, it's just a Perl module that allows for interacting with version control systems in a generic fashion.

I really really want to like this, I really do. The concept is awesome. I like the idea of a modular version-control-system-independent way to deal with patches and get context on them. The dependency on Moose and the performance penalty under mod_cgi is a showstopper for me. I love trying out new technology, but I've also been working as a sysadmin long enough to know that not everyone has the latest technology at their disposal, and we have to be careful about what we throw at people that would either prevent them from ever upgrading Bugzilla again (thus leaving them with vulnerable systems in case of security problems discovered later that don't get patched in older versions) or cause a performance regression running on what they already have. People *CARE* about Bugzilla's performance, a lot. Despite the major strides we've made on performance, it's still one of the number one complaints we get (aside fromthe ugly user interface). Causing a performance regression is show-stopper for me. Comments on CPAN suggest that VCI no longer has a maintainer. Without a maintainer, there's little likelihood that these performance concerns will be addressed. As such, there's not much point in continuing here. We can revisit this later if the situation changes.