Hi Facundo, Try to flush stdout explicitly in the test. The plugin and the compiler may not share the same libc buffers and the compiler only flushes its own.
Cheers, Sylvain On Wednesday, October 7th, 2026 at 4:08 PM, Facundo DomÃnguez via ghc-devs <[email protected]> wrote: > Dear devs, > > I'm a tad puzzled with a test failure on a MR [1] I'm preparing to > allow annotations in simple interfaces (e.g. those produced by > haddock). Before I debug this further, I would like to check the > basics with someone more familiar with GHC on Windows, if possible. > > The failing test checks that simple interfaces do have annotations, > and this test is part of my contribution. The annotations are simply > missing on windows. This test passes on all other platforms. I looked > for conditional compilation that could affect this, but haven't > spotted anything relevant yet. Other plugin and annotation tests seem > to run fine on Windows. > > The change itself only modifies two lines of code, which I paste below > for convenience. > > Any conjectures are welcome. Best, > Facundo > > [1] https://gitlab.haskell.org/ghc/ghc/-/merge_requests/16754 > > diff --git a/compiler/GHC/Iface/Tidy.hs b/compiler/GHC/Iface/Tidy.hs > index b3537514ac4a..2ddde284c2be 100644 > --- a/compiler/GHC/Iface/Tidy.hs > +++ b/compiler/GHC/Iface/Tidy.hs > @@ -172,7 +172,8 @@ mkBootModDetailsTc logger > tcg_fam_insts = fam_insts, > tcg_complete_matches = complete_matches, > tcg_mod = this_mod, > - tcg_default_exports = default_exports > + tcg_default_exports = default_exports, > + tcg_anns = anns > } > = -- This timing isn't terribly useful since the result isn't forced, but > -- the message is useful to locating oneself in the compilation process. > @@ -184,7 +185,7 @@ mkBootModDetailsTc logger > , md_insts = insts' > , md_fam_insts = fam_insts > , md_rules = [] > - , md_anns = [] > + , md_anns = anns > , md_exports = exports > , md_complete_matches = complete_matches > }) > > -- > All views and opinions expressed in this email message are the > author's personal opinions and do not represent those of the > organization or its customers. > _______________________________________________ > ghc-devs mailing list -- [email protected] > To unsubscribe send an email to [email protected] > _______________________________________________ ghc-devs mailing list -- [email protected] To unsubscribe send an email to [email protected]
