scala3 support for slick #223
Conversation
pjfanning
commented
Aug 14, 2023
•
pjfanning
commented
- relates to scala3 support for Slick connector #151
- I don't intend to merge this until Slick 3.5.0 is released
pjfanning
force-pushed
the
scala3-slick
branch
from
475defa
to
ee62b08
Compare
pjfanning
force-pushed
the
scala3-slick
branch
from
ee62b08
to
1c05586
Compare
mdedetrich
force-pushed
the
scala3-slick
branch
from
e5f1f41
to
d52180d
Compare
|
So the slick tests are definitely flaky, I added a commit which seems to have helped (i.e. only failing ~30% of the time). I don't recall if the Slick tests were normally this flaky (don't remember this being the case) and it only seems to be happening on Scala 3. |
|
The Slick tests haven't been all that flaky up until now. |
Yes this is my recollection as well, the flakiness is also only occurring with Slick Scala 3 |
mdedetrich
force-pushed
the
scala3-slick
branch
3 times, most recently
from
7633c0e
to
1c05586
Compare
pjfanning
force-pushed
the
scala3-slick
branch
from
f68f996
to
cca7284
Compare
|
@pjfanning There are MiMa related failures, although tbh we never agreed for MiMa/SemVer on this pekko module and Slick 3.5 is slated for 1.1.x anyways |
pjfanning
force-pushed
the
scala3-slick
branch
from
eaa3676
to
861d6f4
Compare
mdedetrich
force-pushed
the
scala3-slick
branch
4 times, most recently
from
0abbbab
to
b41fad0
Compare
|
@pjfanning I just cleaned up the branch which included some rebases so I would just pull the latest changes ignoring the local ones. Good news is that Slick 3.5.0-M4 seems to work fine with none of the previous problems that we were experiencing. |
mdedetrich
force-pushed
the
scala3-slick
branch
from
cc27dc5
to
1328a0c
Compare
|
So it is being asked if we can publish these changes so that people can test Slick 3.5.0-RC1 before Slick does a full release. Since 1.0.x branch is already setup if we want to do snapshots we can just merge this PR as is, milestone would obviously require a proper vote which I am not against and given what we discussed at #443 (comment) creating a milestone right now (aside from just Slick 3.5.0-RC1) may not be a bad idea since we already did quite a few major dependency updates/changes. @pjfanning @nafg wdyt? |
|
Please don't merge this PR until we make a discussion about whether its appropriate for us as a snapshot |
|
Ideally I think releases should not depend on snapshot/RC dependencies, and the main branch should always be in a release-able state. I guess we can deviate from that if there is a strong reason for it. Alternatively, would it be possible to set up our GitHub Actions so that we can publish a snapshot built off a branch/PR? |
|
I agree that it is too time consuming for us to do releases to test non-ASF jars. If someone really wanted to test this, they could build pekko-connectors-slick jars themselves by doing a git checkout of this PR's branch and running sbt commands. |
|
Replied at slick/slick#2891 (reply in thread) |
slick needs newer logback Update Slick.scala Revert "Update Slick.scala" This reverts commit fc62669.
pjfanning
requested review from
mdedetrich,
He-Pin,
nvollmar,
samueleresca,
Roiocam and
raboof
| @@ -54,6 +51,7 @@ object Slick { | |||
| session: SlickSession, | |||
| query: String, | |||
| mapper: JFunction[SlickRow, T]): Source[T, NotUsed] = { | |||
|
|
|||
There was a problem hiding this comment.
just spreading out the code to make it more readable
