At 19:48 Uhr +0100 14.04.2002, Finlay Dobbie wrote: >On Sunday, April 14, 2002, at 01:14 PM, Martin Costabel wrote: > >>My interpretation of the few cryptic remarks I saw is the following (I >>may be wrong, of course): >> >>Apple's forthcoming release will be based on the "nervous weasel" branch >>of their cvs tree, not on the trunk. In the trunk, they are trying to >>keep up with GNU, so there will eventually be an Apple gcc-3.1 based on >>FSF's released gcc-3.1. The branch, however, was forked sometime in >>March and is (or calls itself) "based on gcc version 3.1 20020105 >>(experimental)". That's why I suspect the "experimental" will still be >>there for quite a while. > >Ah, that sounds quite likely actually, I think I've heard something >along those lines from Stan Shebs on the darwin-development list a >while back, come to think of it. Pfft... Typical Apple behaviour >though: ship a non-final compiler ;-)
Well, one could argue whether that is bad or not. The alternative would be to delay the new compiler even more and wait till the FSF gets around to finalizing GCC 3.1, which is scheduled for mid-april - but you know how it is with schedules. But even once gcc 3.1 is finalized, they can't just ship it! They sti ll have to have a ton of changes atop it, and a compiler which is used to compile the entire operating system and many many applications better be rock stable and proven to generate correct code. Considering the coupld of disasters the GCC followers had to go thru with past 3.x release (which had to be quickly superceded by fixed version due to critical bugs), Apple just can't trust the 3.1 release immediatly. So add at least another month, that'd be mid-may. But it's vital that a broad range of developers get access to the 3.1 tool chains ASAP, so that they can start adapting their code base. In the Linux worlds, they had gcc 3.x for ages now, so many apps are adpated (and if you wrote your C++ code only against gcc 2.95.2, you *will* have to make a lot of adjustments to compile on a proper C++ compiler). So, for once, I don't think this was a bad choice, but rather about the best thing Apple could do (short of releaseing a 3.1 chooltain even earlier <g>). Max -- ----------------------------------------------- Max Horn Software Developer email: <mailto:[EMAIL PROTECTED]> phone: (+49) 6151-494890 _______________________________________________ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
