On 5/15/07, John Cowan <[EMAIL PROTECTED]> wrote:
Brandon Van Every scripsit: > We've had a feature complete, field tested, all but bug free CMake > build for awhile now. It took 1 man year to produce. It is the only > way to produce the MSVC versions of Chicken, and it is preferred for > MinGW also. And for Cygwin, thanks to the libffi support. > It works fine for all the builds on all the platforms, actually, > which is why I want to retire the Autoconf system. Almost. We still have to break the circular dependency that the Linux version (and possibly other Unix versions) think we have.
Which one is that? If you're referring to http://trac.callcc.org/ticket/158I reopened it after Felix had closed it as "wontfix." Having the 2nd invocation of "make install" repeat the build from scratch is not that important a bug, and all evidence says it's a Linux specific bug. This level of defect is not a reason to keep the Autoconf build around. In the real world, build systems occasionally have minor bugs. How many bugs do you think people could report with Autoconf itself if they really really tried? Cheers, Brandon Van Every
_______________________________________________ Chicken-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/chicken-users
