Magnus pushed to branch wip/ci-note-treeless at Glasgow Haskell Compiler / GHC
Commits:
d99f8326 by Cheng Shao at 2025-12-11T19:14:18-05:00
compiler: remove unused CPP code in foreign stub
This patch removes unused CPP code in the generated foreign stub:
- `#define IN_STG_CODE 0` is not needed, since `Rts.h` already
includes this definition
- The `if defined(__cplusplus)` code paths are not needed in the `.c`
file, since we don't generate C++ stubs and don't include C++
headers in our stubs. But it still needs to be present in the `.h`
header since it might be later included into C++ source files.
- - - - -
46c9746f by Cheng Shao at 2025-12-11T19:14:57-05:00
configure: bump LlvmMaxVersion to 22
This commit bumps LlvmMaxVersion to 22; 21.x releases have been
available since Aug 26th, 2025 and there's no regressions with 21.x so
far. This bump is also required for updating fedora image to 43.
- - - - -
495389d3 by Cheng Shao at 2025-12-12T09:01:59+00:00
ci: use treeless fetch for perf notes
This patch improves the ci logic for fetching perf notes by using
treeless fetch
(https://github.blog/open-source/git/get-up-to-speed-with-partial-clone-and-s...),
to avoid downloading all blobs of the perf notes repo at once, and
only fetch the actually required blobs on-demand when needed. This
makes the initial `test-metrics.sh pull` operation much faster, and
also more robust, since we are seeing an increasing rate of 504 errors
in CI when fetching all perf notes at once, which is a major source of
CI flakiness at this point.
Co-authored-by: Codex
- - - - -
3 changed files:
- .gitlab/test-metrics.sh
- compiler/GHC/Driver/CodeOutput.hs
- configure.ac
Changes:
=====================================
.gitlab/test-metrics.sh
=====================================
@@ -17,12 +17,14 @@ fail() {
function pull() {
local ref="refs/notes/$REF"
- # 2023-10-04: `git fetch` started failing, first on Darwin in CI and then on
- # Linux locally, both using git version 2.40.1. See #24055. One workaround is
- # to set a larger http.postBuffer, although this is definitely a workaround.
- # The default should work just fine. The error could be in git, GitLab, or
- # perhaps the networking tube (including all proxies etc) between the two.
- run git -c http.postBuffer=2097152 fetch -f "$NOTES_ORIGIN" "$ref:$ref"
+
+ # Fetch performance notes from a dedicated promisor remote using a
+ # treeless filter, so that individual note blobs are fetched lazily
+ # as needed.
+ git remote add perf-notes "$NOTES_ORIGIN" || true
+ git config fetch.recurseSubmodules false
+ git config remote.perf-notes.partialclonefilter tree:0
+ run git fetch --force perf-notes "$ref:$ref"
echo "perf notes ref $ref is $(git rev-parse $ref)"
}
@@ -81,4 +83,3 @@ case $1 in
pull) pull ;;
*) fail "Invalid mode $1" ;;
esac
-
=====================================
compiler/GHC/Driver/CodeOutput.hs
=====================================
@@ -329,15 +329,8 @@ outputForeignStubs logger tmpfs dflags unit_state mod location stubs
stub_c_file_exists
<- outputForeignStubs_help stub_c stub_c_output_w
- ("#define IN_STG_CODE 0\n" ++
- "#include \n" ++
- rts_includes ++
- ffi_includes ++
- cplusplus_hdr)
- cplusplus_ftr
- -- We're adding the default hc_header to the stub file, but this
- -- isn't really HC code, so we need to define IN_STG_CODE==0 to
- -- avoid the register variables etc. being enabled.
+ (rts_includes ++
+ ffi_includes) ""
return (stub_h_file_exists, if stub_c_file_exists
then Just stub_c
=====================================
configure.ac
=====================================
@@ -526,7 +526,7 @@ AC_SUBST(InstallNameToolCmd)
# versions of LLVM simultaneously, but that stopped working around
# 3.5/3.6 release of LLVM.
LlvmMinVersion=13 # inclusive
-LlvmMaxVersion=21 # not inclusive
+LlvmMaxVersion=22 # not inclusive
AC_SUBST([LlvmMinVersion])
AC_SUBST([LlvmMaxVersion])
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/505caf1d6a41fe339dfec41c9e81a72...
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/505caf1d6a41fe339dfec41c9e81a72...
You're receiving this email because of your account on gitlab.haskell.org.
