Skip to content

Add base dependency for data-default-class - #32

Open
TeofilC wants to merge 1 commit into
mauke:flarefrom
TeofilC:patch-1
Open

TeofilC wants to merge 1 commit into
mauke:flarefrom
TeofilC:patch-1

Conversation

@TeofilC

@TeofilC TeofilC commented Aug 28, 2026

Copy link
Copy Markdown

The next version of GHC requires a dependency on base if any built-ins are used. This package won't compile without adding this dependency.

The next major version of GHC requires a dependency on GHC if any built-ins are used. This package won't compile without adding this dependency.
@TeofilC

TeofilC commented Aug 29, 2026

Copy link
Copy Markdown
Author

See DanBurton/composition#6 for a similar PR for another package that didn't have a base dependency

@mauke

mauke commented Aug 29, 2026

Copy link
Copy Markdown
Owner

I think I'll wait for the next ghc to be released so I can test whether this change really is necessary (and sufficient).

In particular, I think @simonpj is being silly here:

Ah! But you can't know that it "uses no known names"! Suppose GHC desugared f x as (apply f x) where apply is a library function. (Scala does this -- you can overload function application.) Then apply is precisely a known name.

In my case we're talking about an empty module. (Well, mostly empty: It does import/export things, but it contains no declarations.) If GHC needs to inject additional names to implement nothing at all, that seems more like an issue with GHC.

@TeofilC

TeofilC commented Sep 2, 2026

Copy link
Copy Markdown
Author

FWIW it should be possible to test this using your current CI setup, since I think haskell-ci supports testing against unreleased GHCs/head.hackage. Of course I can understand wanting to wait.

I have also added a patch to head.hackage with this change.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants