| |||||||||||||
OLS: Shuttleworth on free software developmentOLS: Shuttleworth on free software developmentPosted Jul 31, 2008 22:08 UTC (Thu) by giraffedata (subscriber, #1954)In reply to: OLS: Shuttleworth on free software development by NAR Parent article: OLS: Shuttleworth on free software development
if the current Fedora, Ubuntu or SuSE includes three different versions of X, the maintainance load on the X developers is three times bigger because they get bugs from three different releases Don't they just largely ignore reports against releases other than current? I mean, a developer tries to apply a bug report to the current release, but if it isn't easy, he tells the reporter to reproduce it in current code. And if the developer creates a fix, he tests it and releases it in the current release. Any support for older releases shipped by packagers has to come from those packagers. Now, where a project actually maintains 3 release streams (i.e. there is a current 1.x, a current 2.x, and a current 3.x), it's probably because it's in the project's best interest to have users of all three at once -- the 1.x is more stable than the 3.x, so you want people to use 1.x if they can and reduce the total bug handling burden.
(Log in to post comments)
OLS: Shuttleworth on free software development Posted Aug 1, 2008 7:14 UTC (Fri) by NAR (subscriber, #1313) [Link] Don't they just largely ignore reports against releases other than current?Probably, and that's pretty sad. Any support for older releases shipped by packagers has to come from those packagers. But I wonder that how many package maintainers are up to this task, especially with bigger software like OpenOffice...
OLS: Shuttleworth on free software development Posted Aug 1, 2008 15:59 UTC (Fri) by giraffedata (subscriber, #1954) [Link] Any support for older releases shipped by packagers has to come from those packagers.But I wonder that how many package maintainers are up to this task, especially with bigger software like OpenOffice... Well, it's not the task you may be imagining. It's very little of debugging and writing code. Rather, it's tracking the fixes in the current release and backporting ones you need, and reproducing problems against the current release and reporting those upstream. So it's more of a user task than a developer task. I think I may see Shuttleworth's point now. If upstream is going to stop maintaining Release N and start maintaining Release N+1, it's nice for downstream if it happens at the same time that downstream stops maintaining its package that includes Release N and starts maintaining one that includes N+1. That way, nobody has to maintain Release N. If and only if all the upstream projects switch at the same time, it's possible for all the downstreams to pick that time to switch as well.
OLS: Shuttleworth on free software development Posted Aug 7, 2008 18:48 UTC (Thu) by renox (subscriber, #23785) [Link] In theory you're right, in practice .. things may be different! To fix an issue in zebra, I had to find myself the patch in quagga (a fork of zebra) and tell the distribution (that we pay quite a lot for support) to integrate it. Maybe other distribs are better, I don't know, but let's just say that I wasn't impressed by this one.
|
|||||||||||||