Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 18f6138a by Simon Jakobi at 2026-06-04T20:20:31-04:00 testsuite: Deduplicate --only test names config.only is assumed to be a set, but supplying --only overwrote it with the (list) argparse result, which can contain duplicates. When a test ran, config.only.remove(name) dropped only the first occurrence, so a duplicated name lingered and was later misreported as a "test not found" framework failure. Store it as a set instead. Fixes #27322 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> - - - - - 9c26952d by Simon Jakobi at 2026-06-07T21:35:09-04:00 testsuite: detect fast bignum via ghc-internal, not removed ghc-bignum The ghc-bignum package was merged into ghc-internal, so the BIGNUM_GMP probe in test.mk ran `ghc-pkg field ghc-bignum exposed-modules`, which fails with "cannot find package ghc-bignum". That error went to stderr and leaked into the captured stderr of every makefile_test, causing spurious [bad stderr] failures across the suite. The probe also silently returned empty, so config.have_fast_bignum was wrongly False even on GMP builds. Probe ghc-internal's extra-libraries for the gmp library instead: the GMP backend module is an other-module (not exposed), but GMP_LIBS adds gmp to extra-libraries only on a GMP build, so this distinguishes the backends. Redirect stderr to keep any future missing-package error off the harness's stderr. This also removes a stale comment as per suggestion from hsyl20. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> - - - - - 7da2b325 by Alan Zimmerman at 2026-06-07T21:35:10-04:00 EPA: Rename Transform.anchorEof to addModuleCommentOrigDeltas This now matches what it actually does. - - - - - 4 changed files: - testsuite/driver/runtests.py - testsuite/mk/test.mk - utils/check-exact/Main.hs - utils/check-exact/Transform.hs Changes: ===================================== testsuite/driver/runtests.py ===================================== @@ -133,7 +133,7 @@ if args.unexpected_output_dir: config.unexpected_output_dir = Path(args.unexpected_output_dir) if args.only: - config.only = args.only + config.only = set(args.only) config.run_only_some_tests = True if args.skip: ===================================== testsuite/mk/test.mk ===================================== @@ -109,9 +109,11 @@ endif HAVE_GDB := $(shell if gdb --version > /dev/null 2> /dev/null; then echo YES; else echo NO; fi) HAVE_READELF := $(shell if readelf --version > /dev/null 2> /dev/null; then echo YES; else echo NO; fi) -# we need a better way to find which backend is selected and if --check flag is -# used -BIGNUM_GMP := $(shell "$(GHC_PKG)" field ghc-bignum exposed-modules | grep GMP) +# Detect whether the fast (GMP) bignum backend is in use. The GMP backend module +# in ghc-internal is hidden, so we look instead for the gmp library it links +# against: GMP_LIBS adds gmp to ghc-internal's extra-libraries only on a GMP +# build. +BIGNUM_GMP := $(shell "$(GHC_PKG)" field ghc-internal extra-libraries 2>/dev/null | grep gmp) ifeq "$(filter thr, $(GhcRTSWays))" "thr" RUNTEST_OPTS += -e config.ghc_with_threaded_rts=True ===================================== utils/check-exact/Main.hs ===================================== @@ -646,7 +646,7 @@ addLocaLDecl3 :: Changer addLocaLDecl3 libdir top = do Right newDecl <- withDynFlags libdir (\df -> parseDecl df "decl" "nn = 2") let - doAddLocal = replaceDecls (anchorEof lp) [parent',d2'] + doAddLocal = replaceDecls (addModuleCommentOrigDeltas lp) [parent',d2'] where lp = top (de1:d2:_) = hsDecls lp @@ -667,7 +667,7 @@ addLocaLDecl4 libdir lp = do Right newDecl <- withDynFlags libdir (\df -> parseDecl df "decl" "nn = 2") Right newSig <- withDynFlags libdir (\df -> parseDecl df "sig" "nn :: Int") let - doAddLocal = replaceDecls (anchorEof lp) (parent':ds) + doAddLocal = replaceDecls (addModuleCommentOrigDeltas lp) (parent':ds) where (parent:ds) = hsDecls (makeDeltaAst lp) @@ -781,7 +781,7 @@ rmDecl3 _libdir lp = do rmDecl4 :: Changer rmDecl4 _libdir lp = do let - doRmDecl = replaceDecls (anchorEof lp) [de1',sd1] + doRmDecl = replaceDecls (addModuleCommentOrigDeltas lp) [de1',sd1] where [de1] = hsDecls lp (de1',Just sd1) = modifyValD (getLocA de1) de1 $ \_m [sd1a,sd2] -> ===================================== utils/check-exact/Transform.hs ===================================== @@ -65,7 +65,7 @@ module Transform , balanceComments , balanceCommentsList , balanceCommentsListA - , anchorEof + , addModuleCommentOrigDeltas -- ** Managing lists, pure functions , captureOrderBinds @@ -724,8 +724,8 @@ balanceSameLineComments (L la (Match anm mctxt pats (GRHSs x grhss lb))) -- --------------------------------------------------------------------- -anchorEof :: ParsedSource -> ParsedSource -anchorEof (L l m@(HsModule (XModulePs an _lo _ _) _mn _exps _imps _decls)) = L l (m { hsmodExt = (hsmodExt m){ hsmodAnn = an' } }) +addModuleCommentOrigDeltas :: ParsedSource -> ParsedSource +addModuleCommentOrigDeltas (L l m@(HsModule (XModulePs an _lo _ _) _mn _exps _imps _decls)) = L l (m { hsmodExt = (hsmodExt m){ hsmodAnn = an' } }) where an' = addCommentOrigDeltasAnn an View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/fcfa08345e2af28169df6963fcb83a6... -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/fcfa08345e2af28169df6963fcb83a6... You're receiving this email because of your account on gitlab.haskell.org.