My reason for sticking with 8.6.5 is that I don't yet trust GHCJS with later GHC versions. Thanks everyone for the pointers and suggestions. I'll look further into this when I have more time. (Alas, it's not my full-time job, so that's probably not until the weekend.)

On Mon, Jun 15, 2020 at 5:11 AM Simon Peyton Jones <simonpj@microsoft.com> wrote:

Is there a reason you want to stick with 8.6.5?

One way to find what you want is to use git-bisect to find the commit that made the change. I think various people have automated setups for this.

Simon

From: ghc-devs <ghc-devs-bounces@haskell.org> On Behalf Of Chris Smith
Sent: 14 June 2020 14:43
To: ghc-devs@haskell.org
Subject: Looking for a recent error reporting change

Hi. In GHC 8.6.5, if I compile a main module that both fails to define `main` and also contains other errors, the other errors are not reported because GHC aborts after realizing that there is no `main`. In GHC 8.10.1, all errors are reported. I'm looking for where this change was made, in the hopes that it's easy to backport to a local branch of GHC 8.6.5. Does anyone know off-hand?