> On 11/5/07, Scott Cantor <[EMAIL PROTECTED]> wrote: > > Anything that's not Windows should stay with autoconf, anything else is > > nuts. > > I don't think supporting cygwin and mingw-msys is nuts.
I think you're misinterpreting me. Last I checked, cygwin supports autotools. I don't know about mingw-msys. What I was trying to say was that the ./runConfigure mess was an absolute nightmare for me as an upstream consumer, and I would not want to see a repeat of that after so much work to undo it. The native Windows build has to be handled separately, and it has been, but unless there's another platform in demand that doesn't support autotools, I really, really believe that any additional platform work needs to be kept in the autotools framework. That's what I was trying to say. If something else is critical to support but doesn't support autotools, then we need a maintainer for that, seems to me. My point at the beginning of this thread was that we should not hold up the 3.0 release if there are minority platforms without project contribution where the autotools build happens to be failing *at the moment*. If that's controversial, so be it; I still feel that way. > Perhaps you should get fedora to have its own rpm and send patched > upstream? Fedora probably works at this moment in time, but that's in fact my position; if people don't step up to support specific platforms that are breaking, leave it to upstream packagers to get it working and then incorporate their feedback for a future release. All I'm saying is that perfection is not achievable when it comes to Unix. > > So I would suggest that specific people are tasked (on the basis of them > > volunteering) to deal with the ongoing maintenance of specific > > platform/compiler combos, and if nobody steps up, you cut it. > > I'll test mingw and cygwin. I can't guarantee I can fix all issues > that pop up, but I can run ./configure and when it breaks try and fix > it and report my findings and fixes. That's the minimum that I was suggesting; if a platform is breaking and somebody says "Hey, I'll give it a try", then you wait and see what they can fix. But if something in the matrix is breaking and nobody has access to it or is willing to test it, I don't see what we can do. Apparently that's controversial, but to me it's common sense. -- Scott --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
