about summary refs log tree commit diff
path: root/compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp
diff options
context:
space:
mode:
authorMatthias Krüger <matthias.krueger@famsik.de>2022-08-07 01:19:32 +0200
committerGitHub <noreply@github.com>2022-08-07 01:19:32 +0200
commit879c17f3f9a460e8bb28becead81111ee2cadd97 (patch)
treeb1ccc1011cd13429f82b23b54b91b0a0f127b890 /compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp
parentaaa054e53c4d43cf047c1b1eeadaedb0eb5a212c (diff)
parenta0e4c1695854db9bf2f4aae3229026608675307d (diff)
downloadrust-879c17f3f9a460e8bb28becead81111ee2cadd97.tar.gz
rust-879c17f3f9a460e8bb28becead81111ee2cadd97.zip
Rollup merge of #100127 - ChrisDenton:remove-init, r=thomcc
Remove Windows function preloading

After `@Mark-Simulacrum` asked me to provide guidance for when optionally imported functions should be preloaded, I realised my justifications were now quite weak. I think the strongest argument that can be made is that it avoids some degree of nondeterminism when calling these functions (in as far as system API calls can be said to be deterministic). However, I don't think that's particularly convincing unless there's a real world use case where it matters. Further discussion with `@thomcc` has strengthened my feeling that preloading isn't really needed.

Note that `WaitOnAddress` needed some adjustment to work without preloading. I opted not to use a macro for this special case as it seemed silly to do so for just one thing (and I don't like macros tbh).
Diffstat (limited to 'compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp')
0 files changed, 0 insertions, 0 deletions