On Mon, 8 Apr 2024 at 10:02, Peter Eisentraut <pe...@eisentraut.org> wrote: > I have tested this with various compilers, and I can confirm that this > shaves off about 5 seconds from the build wall-clock time, which > represents about 10%-20% of the total time. I think this is a good patch.
Great to hear. > Interestingly, if I apply the analogous changes to the makefiles, I > don't get any significant improvements. (Depends on local > circumstances, of course.) But I would still suggest to keep the > makefiles aligned. Attached is a patch that also updates the Makefiles, but indeed I don't get any perf gains there either. On Mon, 8 Apr 2024 at 07:23, Michael Paquier <mich...@paquier.xyz> wrote: > Well, this is also painful with ./configure. So, even if we are going > to move away from it at this point, we still need to support it for a > couple of years. It looks to me that letting the clang folks know > about the situation is the best way forward. I reported the issue to the clang folks: https://github.com/llvm/llvm-project/issues/87973 But even if my patch doesn't help for ./configure, it still seems like a good idea to me to still reduce compile times when using meson while we wait for clang folks to address the issue.
v2-0001-Speed-up-clean-parallel-meson-builds-a-lot.patch
Description: Binary data