-
Notifications
You must be signed in to change notification settings - Fork 101
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
Bump io.takari:takari from 51 to 55 #303
base: master
Are you sure you want to change the base?
Conversation
Lets wait for CI. I suspect removed support for Java 8. |
Yes, spotless-maven-plugin is 11+ so it needed Java 8 hacks (to NOT run) |
I think spotless can also format Scala code, but I have no idea how to enable it. If it is based on |
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.
Looks good to me.
So, this is the "price" of latest Takari parent POM: spotless (palantir) formatted source code, and build will fail if unformatted code is being built: to reformat |
Personally I love this format, whole Maven 3.9 and 4 and all plugins went to this source format as well. POM are formatted as well (element order BUT NOT like "sortpom" that is very bad idea, as dep ordering DOES MATTER). |
Sounds good to me. Formatted code makes maintenance and PR management easier! |
Lets give some time to @headius and others to chime in... |
If you want to enable spotless for scala files, you could use this config file. |
@lefou done, pls review |
polyglot-scala/pom.xml
Outdated
<artifactId>spotless-maven-plugin</artifactId> | ||
<configuration> | ||
<scala> | ||
<scalafmt /> |
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.
I think you need to refer to the .scalafmt.conf
here. Maybe, because it's not in the repository root directory. The format of the files doesn't match the config.
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.
Yup, realized that, but conf file needed change: error was:
Caused by: java.util.NoSuchElementException: version [expected 3.7.3 or 3.7.3]: 3.7.15
polyglot-scala/.scalafmt.conf
Outdated
# Newer versions won't work with Java 8! | ||
version = "3.7.15" |
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.
# Newer versions won't work with Java 8! | |
version = "3.7.15" | |
version = "3.8.1" |
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.
Whatever else is here than "3.7.3" the plugin pukes, with error above, unsure why.
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.
Maybe, we need an explicit version in the plugin config?
<version>3.8.1</version>
But I think we can live with that older version too.
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.
You are right, set for wanted 3.8.1
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.
Looks reasonable.
* are made available under the terms of the Eclipse Public License v1.0 | ||
* which accompanies this distribution, and is available at | ||
* http://www.eclipse.org/legal/epl-v10.html | ||
* Copyright (c) 2012 to original author or authors All rights reserved. This |
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.
This looks a bit ugly and is unexpected (the merged lines). Maybe, because it was formatted without any config before? Not a showstopper, but maybe you could retry with the original files?
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 you were right, re-reformatted with master checkout first and looks better
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.
Very nice. 👏
Worth it for fixing 2-space indent on all the Java files alone. 👍 |
This involves full reformat of the sources using spotless maven pluging. If this commit enters master, it MUST be followed by another one that sets this commit hash as "git blame ignore". No code change, just reformat of Java, Scala and Kotlin source.
Now, Java, Scala and Kotlin sources are formatted. Will let this PR one more day (to get possible feedback from any interested party) and will merge after. |
Updated PR to parent 55. Not merging yet to not stir things with polyglot-ruby, let's see is their issue fixed for start, and then will continue on this path. |
@lefou can you remind me... AFAIR, you insist on Java 8 build time... Due that we are stuck on Takari Lifecycle 2.0.x (as 2.1.x lifts build time requirement to Java 11, but still can produce Java 8 bytecode). Latest Takari Lifecycle brings among other things updated GPG signing, so this "hack" by involving maven-gpg-plugin would become unnecessary as well. Just to be clear, build time we could use Java 11+ (like a matrix of 11, 17, 21, without 8 as today), but the artifacts produced would still be Java 8 bytecode. Simply put "build time Java requirement" would be raised to Java 11. So, do you still insist on Java 8 build time? |
Most discussion happened in #257. In summary. I'm not against dropping Java 8 build time. But since the Scala dialect supports Java 8 runtime, we also should maintain a usable integration test suite. IIUC, if we drop Java 8 build time, we can no longer run Java 8 integration tests unless we add some toolchain setup, which isn't trivial and beyond my maintenance resources.
No. So, if we drop Java 8 build time support, we better also drop Java 8 runtime support for the Scala dialect due to a missing test setup. Or someone contributes an integration test setup for Java 8. Supporting an untested product is pure evil. There is no way to reproduce issues anymore. |
Ok, all interested folks, I am about to "revive" this PR. It will imply also moving to Java 11 or maybe even Java 17 to build (as takari lifecycle is Java 11, and about to move to 17). Still, will try to set up ITs/UTs run on Java 8 w/ similar setup as on MIMA https://github.com/maveniverse/mima (as explained here: https://cstamas.org/blog/2024/09/very-simple-ci-setup/) |
I am good with moving to 11 and also 17 .. probably not newer than that for this project though. And yes .. we should completely drop for Java 8 - build and runtime. And if we go for 17 .. we would also drop build and runtime with any < 16 |
JRuby 9.4.x will continue to be supported until 2026 at least, and it builds on Java 8 with polyglot-maven. Removing runtime support going forward would mean one of two things:
JRuby 10, due out early in 2025, will move to Java 17 or 21 minimum, so the problem will go away... but not for at least another year after that, and probably more like two (9.4 is a well-established release line and we predict users will drag their feet beyond the 2026 EOL). |
No, you misunderstood me: I plan to move takari-lifecycle to Java 17 (see takari/takari-lifecycle#194). That will mean that we should move building of polyglot to Java17 as well (as I want to move past stale lifecycle used here). OTOH, we can and still should produce Java8 bytecode, and I do plan to do it. But, this comes at cost, that we need split "build" vs "tests" (as in ITs), that is somewhat started by me here: |
So the goal is to build + UT on Java 17 or 21, and then have the IT suite run on matrix that includes Java 8 as well. |
@cstamas Ok, then I'm on board. As long as polyglot still supports the use of Java 8 at runtime (build time for a polyglot-using project) then we will not be affected. |
Right, this is only about building polyglot, but the built polyglot itself remains Java 8, as it was. This is merely for "tooling" as takari-lifecycle is already at Java 11, but latest Eclipse JDT update will push it toward Java 17, and polyglot uses really old release of takari-lifecycle |
This involves full reformat of the sources using spotless maven pluging. If this commit enters master, it MUST be followed by another one that sets this commit hash as "git blame ignore".
Changes:
Fixes #262 implicitly, by a minor cleanup among deps