#13433: Segmentation faults in profiled way -------------------------------------+------------------------------------- Reporter: bgamari | Owner: simonmar Type: bug | Status: new Priority: highest | Milestone: 8.2.1 Component: Compiler | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by bgamari): Alright, I have once again reproduced this. Unfortunately I realized that you actually need to cherry-pick a few patches on top of 6ebfbd as it doesn't build on its own. One of these patches fixes a silly typo. The other is my rather crude fix to #13233 (Phab:D3063) ensuring we don't attempt to tick string literals. I'm a bit suspicious of the latter, but the build doesn't build any lint warnings so I've been operating under the assumption that it's safe. Without further ado, here is a full repro, {{{ #!/bin/bash -e git clone git://git.haskell.org/ghc --recursive ghc-T13433 cd ghc-T13433 cat >mk/build.mk <<'EOF' BuildFlavour = prof ifneq "$(BuildFlavour)" "" include mk/flavours/$(BuildFlavour).mk endif GhcStage2HcOpts += -dcore-lint -dcmm-lint define add_mods_flag = $(foreach mod,$(2),$(eval $(basename $(mod))_HC_OPTS += $(1))) endef $(call add_mods_flag,-fprof-auto,$(wildcard compiler/typecheck/*.hs)) STRIP_CMD = : EOF ./boot ./configure make -j8 wget https://ghc.haskell.org/trac/ghc/raw-attachment/ticket/13387/Main.hs inplace/bin/ghc-stage2 -O2 -fforce-recomp Main.hs +RTS -p }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13433#comment:15 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler