-
Notifications
You must be signed in to change notification settings - Fork 8
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support GHC 9.8 and GHC 9.10 #130
Conversation
5d41995
to
c04838f
Compare
Yeah, I can see how I screwed up with |
6b49557
to
e66339f
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seems very uncontroversial to me. Once you got CI working it can be merged from my perspective.
1ef4e3e
to
be19a10
Compare
I suppose this supercedes #110? |
Oh! I've hardly ever worked on this repository and completely forgot to look at existing PR's. My suggestion is to incorporate those commits into this PR, do you agree? I can also do it the other way around but that would require pushing to your fork; it seems simpler to use this branch. |
Only clash-protocols-base needs to depend upon ghc.
be19a10
to
4757ea0
Compare
GHC 9.8 and 9.10 give a lot of Haddock warnings. We used to check the Haddock for all GHC versions, but this is overkill. Just check it on GHC 9.6 only. We used to first invoke `cabal build` and then `cabal build --enable-documentation`. But this rebuilds all dependencies, effectively doubling running time. The `allow-newer` on `circuit-notation` is superfluous. Further details: Haddock for GHC 9.8 and up generate more information about type family instances. However, this also leads to "could not find link destinations for" warnings. Given an `instance C A`, it would seem that when it lists the instance under the Haddock of the definition of `C`, it tries to create a link to the listing of the instance under the Haddock of the definition of `A`, and fails somehow. I suspect a bug in Haddock rather than something we can fix.
4757ea0
to
94e6c31
Compare
Yes, feel free the take my changes. I just wanted to make sure we avoided duplication of work where possible. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seems uncontroversial :)
sed <.ci/cabal.project.local.in >cabal.project.local " | ||
s/__CHECK_HADDOCK__/$check_haddock/" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why not sed -i
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
But then you'd still have to copy the file, right?
If I try to use --in-place
, I end up with something like:
sed .ci/cabal.project.local.in -i -e "
s/__CHECK_HADDOCK__/$check_haddock/"
cp .ci/cabal.project.local.in cabal.project.local
This PR includes the commits of PR #110 (which I initially had not noticed when I opened this PR).
Updates the
circuit-notation
plugin to current master.Protocols.Internal
actually contains quite some orphan instances regardless ofclash-prelude
version now.Rowan or Tim, could you verify I did the right thing in the Fix partial function warnings in PacketStream commit? By the time I got to
makePropPacketArbiter
I was frankly spent and my error message is pretty non-helpful.Funny thing:
head
has a warning butlast
doesn't. So we could just rewrite every use ofhead
tolast . reverse
. For some reason I didn't.