The State Of GCC 4.7.0: Still Months Away
From Jakub's message, the trunk code for GCC 4.7 should be done with state one by the end of October, if the same 4.6 schedule roughly follows. He's called out on various branch maintainers to see if their respective feature work will be ready in time for merging to GCC 4.7 trunk within the next month and a half.
In response to Jelinek's call for finding out about the un-merged branches, Richard Guenther mentioned the bit-field lowering support is not likely to happen. Andrew MacLeod also mentioned that some parts of the transactional memory GCC branch may be ready for merging in October, but not the entire branch. Vladimir Makarov also responded to say that the GCC LRA support is a long project and that it won't be ready for GCC 4.7, maybe by GCC 4.8, but more than likely by GCC 4.9. Jan Hubicka was the latest to respond and says he might have some more LTO updates for this next GCC release. There's no word yet whether Intel's Cilk Plus will be ready for merging in 4.7.
For those interested in the current end-user features of GNU Compiler Collection 4.7, see the changes page.
As of the status update, there are six P1 regressions, ten new P2 regressions (95 total), and 56 new P3 regressions (59 total).
6 Comments
