-
Notifications
You must be signed in to change notification settings - Fork 374
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
Switch to data-default
from data-default-class
#1852
base: master
Are you sure you want to change the base?
Switch to data-default
from data-default-class
#1852
Conversation
data-default
instead of data-default-class
. Fixed https://github.com/yesodweb/yesod/issues/1850
data-default
instead of data-default-class
. Fixed https://github.com/yesodweb/yesod/issues/1850data-default
from data-default-class
2085a24
to
6251afb
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.
@Oleksandr-Zhabenko FYI, this is failing to build.
@jezen this is due to the following. Stackage still is missing the newer version that will completely resolve the issue. It is present only on Hackage. Since the files changed are all .cabal files, I thought that it would update the needed versions from the Hackage, but it does not. It looks like it is needed to use conditional builds in the .cabal files and CPP extensions in the source files to check the package versions. |
But if we use CPP and conditional .cabal files then the issue #1850 is still actual, because hledger still uses data-default package, which is preferred in the future, instead of data-default-class. |
As thielema said, this change is for the future when the data-default-0.8 is available on Stackage. |
But wait, stack has functionality to deal with such cases. |
These two commits |
cf48ba0
to
2a0fb19
Compare
There are a couple of build failures on ubuntu-latest, --resolver lts-22. One is like this.
The other is like this.
Do you think you can fix these? Happy to support you on this. |
The changes needed to be done for the correction of the first one is related to tls-2.1.2 package that should be used instead of tls-1.8. But this leads to the following error: connection > /tmp/stack-233a16a5e48d1819/connection-0.3.1/Network/Connection.hs:100:33: error: And the connection-0.3.1 is the last available version, and their source repository is archived on GitHub by the author. |
As for the CPP preprocessing macros, I think it is better to delete the ifdef part and use just the following if-else. |
I have just tried to resolve the non-building cases, this led to updating many dependencies in the stack.yaml file on the project level. But still, the connection package is not compiled because it has no appropriate version, it is needed to contact @vincenthz to provide some workaround for this or to fork his archived repository and then try to add changes from there to the Hackage package connection. |
Can you switch to this maintained fork instead: https://hackage.haskell.org/package/crypton-connection ? |
It seems to be resolved. I will now squash commits to have just one final with the needed changes. |
7c550dd
to
5f6db37
Compare
I see that there are still some CI errors. Probably we can remove the old resolver. I think having CI for the last 3 GHC major versions should be good enough. |
I have the same opinion. The fails are for the lts-16 resolver only, and are connected with base-4.13* and basement-0.14* series instead of the provided newer versions. |
Fixed issue with dependencies for data-default branch Resolved non-building cases using crypton-connection Resolved non-building cases by dropping support to lts-16
9eb8563
to
ac33267
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.
Now it builds.
Switched to data-default instead of data-default-class. Fixed #1850