2011/8/10 Mook <mookgcc+gm...@gmail.com>: > On 8/8/2011 8:02 AM, Ruben Van Boxem wrote: >> Dear mingw-w64 developers, > > First, welcome in your new packager hat! Except not quite, because > you've been wearing a similar hat for a while already ;)
True, but this has a different goal of course (lots more platforms to build for, and fixed versions to appease the masses) > > (I'm just going to ruthlessly snip all the bits I agree with, so please > read this as violently agreeing ;) ) > >> - What we agreed on about versioning: >> --> mingw-w64 should adopt a semi-rolling release model. It's >> source only, so (Linux) packagers should just pick a (release) branch >> and use its latest revision. >> --> binutils: latest trunk is the only sensible version >> --> GCC: latest stable versions, meaning the latest released version. >> --> Each mingw-w64 toolchain release will happen for each GCC >> release, and this over (currently) the 4.5 and 4.6 branches. 4.4 is >> not x86_64-w64-mingw32 ready yet, so it's quite senseless to use that. >> Each new GCC minor version will ideally be accompanied by an update on >> mingw-w64 release side. > This will fit well for some distros (Arch), not so much for others > (RedHat, should we ever get there). But the opposite way is... well, > opposite, and I'm happy to live with anything. > >> Unless there's a way to cross-compile from linux to mac, I'd need >> access to build machinery. Maybe it would be best to get the releases >> set up on some common infrastructure anyways. > Yeah; there were old attempts to convince buildbot to do > manually-triggered builds as well (optionally from a given tarball URL), > that might be useful? That would be very helpful, in an ideal case it would build for all platforms from exactly the same source. > >> In conclusion, the releases will closely resemble my latest Personal >> builds, which isn't a coincidence ;-) > Hey, as long as it works, and hopefully repeatable :) > >> Any constructive thoughts are welcome, and help in getting the >> machinery set up. In my eyes, this would mean making my scripts >> runnable on some build infrastructure, and expanding them to work for >> Cygwin/Mac, hopefully without much trouble. > If msys works, it's probably not going to be too bad (just remember that > Mac looks more like BSD in some ways). I'm torn between cleaning up the makefile (which may have merit) or replacing it altogether (due to all the old cruft, as you so clearly put it). There is no plan to support GCC 4.4 (due to heavy patching, and the fact that 4.5 and 4.6 work much better), so the old cruft is just that. How does the buildbot work, *exactly*, when I ask it to do a mingw-x86-x86_64 build for example (for simplicity, assume I used the forcesrc page)? This would really help my understanding of it :) Ruben ------------------------------------------------------------------------------ uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev _______________________________________________ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
