I'm tentatively satisfied with the state of the tree for 9.2 in commit cf3024a8. Cross-builds still aren't perfect and liarc can't build itself (or cross-build anything else) in parallel, but it's good enough for now: builds from 9.1.1 and works. Let me know if there are any outstanding issues.
Someone else should try building it in a few configurations just in case I broke something I haven't realized, but it's a lot better now than it was 48 hours ago. Every configuration I've tried so far passes all tests: i386 from 9.1.1/i386 amd64 from 9.1.1/amd64 svm1 from 9.1.1/amd64 (with `--enable-cross-compiling', `make all-native') liarc from 9.1.1/amd64 (with `make all-native') liarc from master/liarc svm1 from master/svm1 (with `make all-native') amd64 from master/svm1 (with `--enable-cross-compiling') Every one of these builds took under an hour on my laptop, which was not otherwise idle. Every one except liarc from master/liarc used the all-native target and took under twenty minutes; liarc from master/liarc took just under an hour, but its build is serial. The slowest part was actually running the tests for svm1 -- it takes over an hour to run them. _______________________________________________ MIT-Scheme-devel mailing list MIT-Scheme-devel@gnu.org https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
