[Git][ghc/ghc][wip/23944] driver: Only run a dynamic-too pipeline if object files are going to be generated
Zubin (@wz1000)
gitlab at gitlab.haskell.org
Wed Dec 6 11:47:27 UTC 2023
Zubin pushed to branch wip/23944 at Glasgow Haskell Compiler / GHC
Commits:
2b69d69e by Zubin Duggal at 2023-12-06T17:17:20+05:30
driver: Only run a dynamic-too pipeline if object files are going to be generated
Otherwise we run into a panic in hscMaybeWriteIface: "Unexpected DT_Dyn state
when writing simple interface" when dynamic-too is enabled
We could remove the panic and just write the interface even if the state is `DT_Dyn`,
but it seems pointless to run the pipeline twice when `hscMaybeWriteIface` is already
designed to write both `hi` and `dyn_hi` files if dynamic-too is enabled.
Fixes #23944.
- - - - -
1 changed file:
- compiler/GHC/Driver/Pipeline.hs
Changes:
=====================================
compiler/GHC/Driver/Pipeline.hs
=====================================
@@ -788,7 +788,15 @@ hscBackendPipeline pipe_env hsc_env mod_sum result =
if backendGeneratesCode (backend (hsc_dflags hsc_env)) then
do
res <- hscGenBackendPipeline pipe_env hsc_env mod_sum result
- when (gopt Opt_BuildDynamicToo (hsc_dflags hsc_env)) $ do
+ -- Only run dynamic-too if the backend generates object files
+ -- See Note [Writing interface files]
+ -- If we are writing a simple interface (not . backendWritesFiles), then
+ -- hscMaybeWriteIface in the regular pipeline will write both the hi and
+ -- dyn_hi files. This way we can avoid running the pipeline twice and
+ -- generating a duplicate linkable.
+ -- We must not run the backend a second time with `dynamicNow` enable because
+ -- all the work has already been done in the first pipeline.
+ when (gopt Opt_BuildDynamicToo (hsc_dflags hsc_env) && backendWritesFiles (backend (hsc_dflags hsc_env)) ) $ do
let dflags' = setDynamicNow (hsc_dflags hsc_env) -- set "dynamicNow"
() <$ hscGenBackendPipeline pipe_env (hscSetFlags dflags' hsc_env) mod_sum result
return res
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2b69d69e56a8e6e1456f29bdda357ee14e3bc77c
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/2b69d69e56a8e6e1456f29bdda357ee14e3bc77c
You're receiving this email because of your account on gitlab.haskell.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20231206/cf1d0eff/attachment-0001.html>
More information about the ghc-commits
mailing list