Quoting Alan Hourihane <[EMAIL PROTECTED]>: > > Sorry, this is not how it goes. We won't be willing > > to adopt anything blindly. There is a German saying > > applying here: > > 'Never buy a cat in the bag!' > > > > 1. First there should be a proposal > > 2. Secondly there should be a test implementation > > as proof of concept we can work with and see > > how well it goes. > > 4. Thirdly this should be evaluated > > - if we think it is usable at all. > > - what modifications we would like to see. > > 5. The modified system needs to get retested and reevaluated. > > 6. This is the earliest stage we can talk about a more or > > less formal policy. > > > > Up to now it is not even clear who should be able to > > submit to this bug tracking system: > > Should it be internal only? > > Should only projects like GNOME, KDE etc and > > distributions like RedHat, SuSE, Mandrake be able to > > submit bugs? > > Or should it be open to the general public? > > I agree with your points here Egbert.
I believe such possible bug track system should at least: o Be very lighweight, I don't want to wait 30+ seconds for every small iteration (for example rebuilding the interface based on a selected list entry) while doing a query in a server in the other side of the globe. o Only allow submits if at least a clear description of how to reproduce the problem is present. Also, some padronized way to query the hardware, configuration, os, and log files must exist and be attached to the pr. Bugs reports with an optional proposed patch would have higher precedence. o Require a valid bug reporter email. Needless to say, the major interest of XFree86 developers is to have a very stable implementation, but since a lot (most) work is done as a volunteer effort, just filling bug reports won't be enough, and saying that a XFree86 developer/contributor is lazy or unresponsible because he cannot handle the 300+ bogus assigned prs won't help either. > Someone needs to step forward to collate these ideas and put the > initial > system together. We seem to have people stepping forward with ideas > and a few who claim that they'll do the initial work on setting up, > but nothing ever happens. > > My own personal opinion lies with the fact that if this is setup, who > will have the staying power to maintain it. Most of the XFree86 Core > Team > have been around a long time and seen various developers come and go > around > the project. It's obvious to me that some developers will use it, and > others may not, or even may not all the time. Therefore that 'team' > who > has to maintain it need to be able to stick around for it to be of > long term use. If these 'bug' maintainers come and go, I can see the > whole > process collapsing into one big heap of unusable data. Even if a bug report system is not done at this "round", I would at least like to see some work about categorizing XFree86 development/ maintenance areas, like: xserver, driver, design, extensions, documentation, libraries, xtest, security, i18n, etc; and the various subdivisions. > There's been several conversations of this in the past and nothing > fruitful has come of it just because the XFree86 Core Team don't push > for it, and so others see that as counter-productive. Don't. Someone > needs to put their own effort in. > > Even if the XFree86 Core Team don't participate in this bug database > it should in no way effect it's usefulness to those who want one. If > someone > enters a bug report, then it's perfectly possible for someone > completely > new to the project to fix the bug and submit a patch to XFree86 > without > any involvement of the Core Team, apart from actually committing the > fix, > then enter onto the bug database 'Fixed - appending commit' in the bug > report. > > Step forward those who want it and start doing it. If it can be shown > to be productive then there's something more to discuss, but we're > covering the same ground again and again. > > Get going on it and prove it works, but don't expect everyone to use > it > and live with that fact - it's open source and you can't force people > into a way of working.... I would like to see some sort of bug tracking system, but only after it has been extensively tested, and only if it proves to be really useful. In the several years I have been involved with XFree86 (some times I got a bit "distant", must agree) what I see is that developers are working in specific/specialized areas, while David Dawes almost alone handles patches and bug reports at "random" parts of XFree86, and also does a large amount of specific/specialized work, so, he is the best person to evaluate it. Hope that after this thread finishes, something useful remains... Paulo _______________________________________________ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
