dep: bump sbt to 1.7.1 #15728
dep: bump sbt to 1.7.1 #15728
Conversation
ckipp01
force-pushed
the
sbtBump
branch
3 times, most recently
from
262769f to
70da34a
Compare
Looking at the [commit](scala@fa063b6) that added 1.6.2 I believe I have everything updated that needs to be. I am updating this because it's needed for a test that I'm adding in scala#15565, but thought it'd be best to do this bump in a separate pr.
This is needed due to some scalablytyped stuff that changed in fs2. Without it, you'll get an error when updating the sbt 1.7.x that scalably typed doesn't support 1.7.x
This also allows us to get rid of the fnGen stuff and simplify the sbtTestCommand and sbtPublishCommand
If not you'll have issues with http4s tests since http4s isn't using the 1.x version of munit yet
We currently have a mess of different versions and because of it cats-effect is using something deprecated in cats. However we can't bump everything due to other issues, so for a temporary fix try turning off xfatal for htpp4s.
|
I think this might be fixed by backing up |
Without doing this it was getting stuck in CI
|
Whoa, what a ride, that was way more involved than I thought it'd be. To give a little recap of why I did this since I ended up making way more changes than originally planned. While working on #15565 which relates to #14904 I needed to bump the sbt version so that our sbt testing would take into account the newer zinc that is included in sbt 1.7.x. So I opened this assuming it'd be a quick bump. Once opened, the fork of fs2 we were using wouldn't work with 1.7.x because of scalablyTyped. In newer version of fs2 that direct dependency was removed in order to make maintenance easier. However, once I bumped fs2 that opened a can of works as you could see. There were also some issues scala8-java-compat and a couple other small things in various places. I originally thought I could just bump everything to the latest, but obviously that didn't work, especially because of version mismatches. For example munit head is really the 1.x milestone series which some in here are using (like cats), but not others. So some things are locked on certain versions. I also noticed while doing this that the majority of the forks are crazy out of date. And some had changes in them that weren't upstreamed. I did a couple prs to upstream some stuff but I'll probably do a couple more in the future and try to sync this all better. I'll start a discussion around this because the current process for the community build is quite difficult and might benefit from being more up to date. Story for another time.
Thank you so much for this. I was just about to start bumping other things in an attempt to fix this, but this was way easier. |
I had tried bumping |
* dep: update fs2 submodule This is needed due to some scalablytyped stuff that changed in fs2. Without it, you'll get an error when updating the sbt 1.7.x that scalably typed doesn't support 1.7.x * dep: sync with latest staging http4s * dep: sync with latest scala-java8-compat This also allows us to get rid of the fnGen stuff and simplify the sbtTestCommand and sbtPublishCommand * dep: sync with latest cats in staging * dep: sync with latest munit-cats-effect * dep: sync with latest discipline and discipline-munit * dep: lock munit at 0.7.29 If not you'll have issues with http4s tests since http4s isn't using the 1.x version of munit yet * disable `tlFatalWarnings` in http4s * dep: lock scalacheck at 1.16.0 Without doing this it was getting stuck in CI
* dep: update fs2 submodule This is needed due to some scalablytyped stuff that changed in fs2. Without it, you'll get an error when updating the sbt 1.7.x that scalably typed doesn't support 1.7.x * dep: sync with latest staging http4s * dep: sync with latest scala-java8-compat This also allows us to get rid of the fnGen stuff and simplify the sbtTestCommand and sbtPublishCommand * dep: sync with latest cats in staging * dep: sync with latest munit-cats-effect * dep: sync with latest discipline and discipline-munit * dep: lock munit at 0.7.29 If not you'll have issues with http4s tests since http4s isn't using the 1.x version of munit yet * disable `tlFatalWarnings` in http4s * dep: lock scalacheck at 1.16.0 Without doing this it was getting stuck in CI
Looking at the
commit
that added 1.6.2 I believe I have everything updated that needs to be. I
am updating this because it's needed for a test that I'm adding in #15565,
but thought it'd be best to do this bump in a separate pr.