Closed ad-si closed 4 years ago
Which GHC version and ghc-lib-parser version were used for the build?
On Sat, Jan 11, 2020, 16:27 Adrian Sieber notifications@github.com wrote:
Tried to build it on macOS with stack and got following error:
[39 of 41] Compiling Ormolu.Parser
/Users/adrian/Tests/ormolu/src/Ormolu/Parser.hs:194:16: error:
• No instance for (Bounded Extension) arising from a use of ‘minBound’ • In the expression: minBound In the expression: [minBound .. maxBound] In an equation for ‘allExts’: allExts = [minBound .. maxBound] |
194 | allExts = [minBound .. maxBound]
| ^^^^^^^^
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/tweag/ormolu/issues/496?email_source=notifications&email_token=AALT42UHEZFJ3UERGIECUJTQ5HCNTA5CNFSM4KFR76HKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IFQMEUQ, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALT42SFHEG4VL56Q2JLKQ3Q5HCNTANCNFSM4KFR76HA .
Perhaps we should go back to pinning the ghc-lib-parser version.
On Sat, Jan 11, 2020, 17:16 Artyom Kazak artyom@artyom.me wrote:
Which GHC version and ghc-lib-parser version were used for the build?
On Sat, Jan 11, 2020, 16:27 Adrian Sieber notifications@github.com wrote:
Tried to build it on macOS with stack and got following error:
[39 of 41] Compiling Ormolu.Parser
/Users/adrian/Tests/ormolu/src/Ormolu/Parser.hs:194:16: error:
• No instance for (Bounded Extension) arising from a use of ‘minBound’ • In the expression: minBound In the expression: [minBound .. maxBound] In an equation for ‘allExts’: allExts = [minBound .. maxBound] |
194 | allExts = [minBound .. maxBound]
| ^^^^^^^^
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/tweag/ormolu/issues/496?email_source=notifications&email_token=AALT42UHEZFJ3UERGIECUJTQ5HCNTA5CNFSM4KFR76HKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IFQMEUQ, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALT42SFHEG4VL56Q2JLKQ3Q5HCNTANCNFSM4KFR76HA .
I was using the stack lts-14.3 as described in your documentation, but with 14.20 and 13.19 it didn't work either…
Yeah, the Bounded instance only appeared in 8.8.1 and lts-14.3 includes 8.8.0.
We would either have to add a more restrictive lower bound, or go via Enum instead of Bounded if we want to keep using the ghc-lib-parser provided by Stackage.
On Sat, Jan 11, 2020, 17:22 Adrian Sieber notifications@github.com wrote:
I was using the stack lts-14.3 as described in your documentation, but with 14.20 and 13.19 it didn't work either…
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/tweag/ormolu/issues/496?email_source=notifications&email_token=AALT42WQSDFV33MYLUYZOLDQ5HI2NA5CNFSM4KFR76HKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIWC7JI#issuecomment-573321125, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALT42UK4VHMB5Q6NBJEBU3Q5HI2NANCNFSM4KFR76HA .
@mrkkrp which one should it be? Do you care about sticking to Stackage versions or not?
The lower bound seems to be correct in both master and the released version on Hackage:
ghc-lib-parser >= 8.8.1 && < 8.8.2
I do not understand how it could build with 8.8.0 given these bounds.
Here is what I use:
resolver: lts-14.03
packages:
- '.'
extra-deps:
- ghc-lib-parser-8.8.1@sha256:81dc4ed6ae6b7ccfda1e62e3c2d9a8600f8ecf45e4237d4dc099ccd4b74db180
Closing, since currently we have correct lower bound, unless I'm missing something here.
Tried to build it on macOS with stack and got following error: