Unless I’m missing a build-script flag, it seems to me that compiling the Swift stdlib with the unoptimized debug swift compiler takes about 15 minutes on a fast machine.
I am assuming that you mean a debug swift compiler building an optimized stdlib?
I’m building the debug swift compiler via ./utils/build-script -r —debug-swift. I assume, perhaps wrongly, that implies a debug stdlib.
Other than forcing the type checker to be optimized, what if any tricks can I use to building the stdlib faster with the debug compiler? Is there a way to tell Clang to enable the inliner and only the inliner during -O0 builds? I have an anecdotal experiment that suggests that this would yield appreciably faster Swift stdlib builds with the debug compiler (and selfishly speaking, I can tolerate the minor impact on debugging that inlining does to otherwise unoptimized code).
Are building LLVM in release + Swift in debug? I.e.:
—release-debuginfo --debug-swift --force-optimized-typechecker
Yes, with the exception that I cannot use —force-optimized-typechecker because I’m hacking on the type checker. Otherwise, this is what I’m doing to make debug builds go as fast as possible:
--llvm-targets-to-build X86 \
--skip-ios --skip-tvos --skip-watchos \
--skip-build-benchmarks true \
--build-swift-static-stdlib false \
--build-swift-static-sdk-overlay false \
--build-swift-dynamic-sdk-overlay false \
--build-swift-stdlib-unittest-extra false \
--extra-cmake-options \\-DCMAKE_CXX_FLAGS=-Werror=switch \
On Aug 6, 2017, at 16:16, Michael Gottesman <email@example.com> wrote:
On Aug 6, 2017, at 11:11 AM, David Zarzycki via swift-dev <firstname.lastname@example.org> wrote:
 – If one force inlines LLVM’s casting logic and associated callbacks (like classof() and getKind()), then the Swift stdlib builds 18% faster on my machine with the debug Swift compiler. One can imagine how much faster the whole stdlib would compile if all trivial functions were inlined automatically.
swift-dev mailing list