#3693: Show stack traces -------------------------------------+------------------------------------ Reporter: jpet | Owner: Tarrasch Type: feature request | Status: new Priority: normal | Milestone: 7.10.1 Component: Runtime System | Version: 6.10.4 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by scpmw): Status update: I have now finished polishing up the initial source-note patches for tracking source code relations throughout the Core and Stg stages: https://github.com/scpmw/ghc/commits/profiling-review2 This is in many ways the most tricky part of the whole endeavour, as it touches pretty much every single pass that does something interesting with Core. It also extends the whole tickish framework a bit, which means cost- centres, HPC ticks and breakpoints have had some of their code rebuilt. On the other hand, it now passes the whole testsuite (minus T4891 and T5435_dyn_asm, which fail on master for me), and I tested that '-g' causes no extra allocations across nofib. I *think* it should be merge-ready for 7.9. Notes for review (Simon?): - I tried to be chatty with the comments and spell out exactly how I understand the properties of tickishs. Some new tickish flags and lots of documentation in `CoreSyn.lhs`. - Given that we now support multiple ticks in the tree at the same time, I had to give `mkTick` a rewrite. - The change to `cheapEqExpr` is a bit awkward, but it was the only cheap way I could think of for making it look through ticks. - Small changes with their own commit are optional changes or refactoring. Next up will the Cmm support, which so far still needs work on the CmmContext replacement. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/3693#comment:58 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler