-
Notifications
You must be signed in to change notification settings - Fork 206
Drop 8.2.1 support? #1249
Comments
I've just checked a readme file and haven't found a section about older GHCs support. I thought it was already supported only on some old branch / release? |
@Anrock that is version 8.0.2 Right now 8.2.1 is the odd one out, as it cannot accept an up to date hlint, for example, because of dependencies. And it just started failing on appveyor for no good reason |
@Anrock for some reason github believes you are in GMT+2, so thinks your posts are in 6 hours time. Or maybe just matches my time zone. So for me the threading of our conversation is broken. |
If the matter is Data.Semigroup and some GHC API #ifdefs we have anyway, I'm against the dropping. CI is parallel anyway. If you want to drop a few CI jobs, it's better to drop some minor versions. Oh, and while I have rights to commit, I don't. So treat this as a user voice. In both C++, Perl or Node.JS old versions are supported pretty well by IDEs because it's what production needs. As for GHC, I have some code using 7.4 because it supported Windows DLLs. And 8.2 is pretty modern, yesterday I saw a question on a Haskell gitter channel about 8.0.2. |
Threading is definitely broken, hehe |
How realistic is it that users are stuck on GHC 8.2.1, without being able to use GHC 8.2.2? iirc, 8.2.1 was basically useless in any case. |
I don't mind dropping 8.2.1 if it's broken and there's unbroken 8.2.x we still support. After all, we have a similar situation with 8.6.3 broken on Windows |
I don't think it is realistic to say people are stuck in 8.2.1 without being able to bump GHC to the next patch version. |
Ok, so next release we will drop 8.2.1 |
We're supporting quite a lot of builds right now, and 8.2.1 has been around for a while. If there is not a lot of interest in keeping it on then we should remove it make CI easier.
It'll also allow us to clean up all those CPP pragmas where we need to explicitly import Data.Semigroup
The text was updated successfully, but these errors were encountered: