mstorsjo wrote: Overall, thanks for the PR - this looks very exciting to me!
First off, I think it'd be good to set the scope for this. This is one of the historical Windows architectures, supported in early versions of NT and CE. As such, it probably has mostly hobbyist value at this point. However that shouldn't really be an issue for LLVM - I don't foresee this to add very much maintainance burden (as this isn't a new target per se, but mostly scattered COFF specific conditions), and is quite interesting for Windows target completeness. The main extra "burden" may be that people interested in testing all Windows configurations now should enable another target backend. As for getting changes properly reviewed and merged; within LLVM, each PR, once merged, results in one commit. So for a longer patch series like this, one would need to do some sort of set of stacked PRs. (Or as a simpler version, one can make one branch per commit, i.e. one for commit 1, one for commits 1+2, one for commits 1+2+3 etc, and just point out the dependencies in each PR.) In practice, I guess we can do some sort of pre-review within this PR, and for the commits that look fine, you can file individual PRs, avoiding needing to have 13 outstanding stacked PRs at once. CCing a couple people who may be interested in commenting on things, @rnk @zmodem @cjacek @efriedma-quic. https://github.com/llvm/llvm-project/pull/107744 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
