Stream: t-compiler/help

Topic: Diagnostics reported against HIR


Taylor Cramer (Apr 24 2019 at 21:57, on Zulip):

Seems like a stretch, but is there a debug mode in which I can see typechecking errors pointed at a pretty-printed HIR rather than the original source spans? Some desugaring code I wrote is getting type mismatch errors, but it's hard to tell which part of the code they're coming from

varkor (Apr 24 2019 at 22:01, on Zulip):

@Taylor Cramer: I wonder if you could use -Zunpretty=hir and then try compiling the output

Taylor Cramer (Apr 24 2019 at 22:03, on Zulip):

@varkor unfortunately that doesn't work

Taylor Cramer (Apr 24 2019 at 22:03, on Zulip):

@varkor in my case, the HIR compiles just fine XD

Taylor Cramer (Apr 24 2019 at 22:04, on Zulip):

so it's some extra bit of metadata I'm setting that's causing it to fail

varkor (Apr 24 2019 at 22:04, on Zulip):

ah :/

davidtwco (Apr 25 2019 at 05:50, on Zulip):

I would love it if we had something like that.

oli (Apr 25 2019 at 08:49, on Zulip):

cc @WG-diagnostics we could probably add a mode where we don't highlight code spans but dump the pretty printed HIR (all levels of expansion even) of whatever span is offending the compiler

Last update: Nov 11 2019 at 22:40UTC