CI: https://gitlab.com/wildfyre/lib/pipelines/58137002

• Prototype of .gitlab-ci.yml
• Clone the submodule with HTTPS instead of SSH (no keys on CI servers)
• Fix parallel tests when there's only 1 available core
• Dockerfile with Java & Python
• What is running on port 8000 in the container?
• Removed after_script
• Cache Gradle dependencies
• Remove pointless lsof
• Cache wheels
• Add virtualenv, PostGresQL, Git, OpenJDK sources, GCC to the container
• Assume API is not running when lsof is missing
• Do not run tests with static analysis
• Cleaning the Dockerfile
• # Ceci est la combinaison de 2 commits.
• Better cache for CI
• Reordering dependencies for clarity
• Pip cache, reports as artifacts, Gradle without daemon
• Run Dokka alongside the unit tests, to speed up builds
• Gradle cache
• Fix query order (posts are not always in the same order)
• Update the API
• Upload JUnit results as artifacts
• Test the Gradle cache before running the scripts
• Documentation in .gitlab-ci.yml
• Upload the JUnit results
• Do not delete the file db.sqlite3 if it doesn't exists (to have a cleaner log output)
• Rename .java to .kt
• Converting Internal to Kotlin

CI: https://gitlab.com/wildfyre/lib/pipelines/58135786

• Prototype of .gitlab-ci.yml
• Clone the submodule with HTTPS instead of SSH (no keys on CI servers)
• Fix parallel tests when there's only 1 available core
• Dockerfile with Java & Python
• What is running on port 8000 in the container?
• Removed after_script
• Cache Gradle dependencies
• Remove pointless lsof
• Cache wheels
• Add virtualenv, PostGresQL, Git, OpenJDK sources, GCC to the container
• Assume API is not running when lsof is missing
• Do not run tests with static analysis
• Cleaning the Dockerfile
• # Ceci est la combinaison de 2 commits.
• Better cache for CI
• Reordering dependencies for clarity
• Pip cache, reports as artifacts, Gradle without daemon
• Run Dokka alongside the unit tests, to speed up builds
• Gradle cache
• Fix query order (posts are not always in the same order)
• Update the API
• Upload JUnit results as artifacts
• Test the Gradle cache before running the scripts
• Documentation in .gitlab-ci.yml
• Upload the JUnit results
• Do not delete the file db.sqlite3 if it doesn't exists (to have a cleaner log output)
• Moving User from Java to Kotlin
• Rename .java to .kt
• Moving Users from Java to Kotlin
• Explaining why the let is ?.let and not !!.let & fixing broken Dokka link
• Rename .java to .kt
• Converting LoggedUser to Kotlin
• Added JsonValue.minimalString & JsonValue.prettyString
• Adding LoggedUser.postsList so that Java code can access the posts
• IssueInTransferException is now able to report on messages that do not parse to JSON
• Better Kotlin integration

• Updating copyrights
• Moving User from Java to Kotlin
• Rename .java to .kt
• Moving Users from Java to Kotlin
• Explaining why the let is ?.let and not !!.let & fixing broken Dokka link
• Rename .java to .kt
• Converting LoggedUser to Kotlin
• Added JsonValue.minimalString & JsonValue.prettyString
• Adding LoggedUser.postsList so that Java code can access the posts
• No more parallel unit tests (the testing server can't handle them)
• IssueInTransferException is now able to report on messages that do not parse to JSON
• Better Kotlin integration

• Updating copyrights
• Rename .java to .kt
• Converting Internal to Kotlin

There are only 2 files in the diff at that point

After updating the API to master, the error doesn't appear anymore. I'm not sure why that is. Logs on GitLab

About the testing part, the current CI script I have for the lib consistently raises it, that could be used for testing. Once this is landed I'll update the API version and see if it still throws the exception

Can you explain why this is important? Also, the dependencies are duplicated now (implementation and fatJar), it this really needed?

@Pregenun
Sorry if it wasn't clear, what I meant was that anyone would have 2 "virtual" areas: "world" which is just everything without sorting, and "subscriptions" that would be every area the user is interested in. Because the posts that are very popular would spread to neighbor areas, you would discover areas that are similar to the ones you know that way. By default, the user would be subscribed to the main ones, but they would slowly discover other ones.

Just tried to apply it against master and it was rejected:

If you remember around where it was from (eg. around which commit on the master branch), I should be able to apply it :)

Sorry for the unclear explanation. Because of the problems with Arc and the time it wastes for us, we will stop using arc unit completely, and instead will implement CI via the GitLab mirror.

This diff originates from ac4b3f808ce91d2f252fb3af08d980e6d8f88b5d which I do not have. I cannot land it.

This diff originates from beb6067caf9de614852fd499083d28871c270d43 which I do not have. I cannot land it.

This is not needed anymore

This is not needed anymore

GitLab too confirms that the diff shown by @Info-Screen is buggy: RequestTest.kt and UserTest.java

• Trying to update the diff to see if the display bug is fixed

• Rename .java to .kt
• Converting Internal to Kotlin

That's weird because the files do have the file definitely has the expected year: https://phabricator-wildfyre-net.info-screen-usercontent.me/file/data/mrntat352z5t7wakas3l/PHID-FILE-me5u4evgkfoqlch3pkrh/src_test_java_net_wildfyre_users_UserTest.java

In D212#5686, @Info-Screen wrote:

Why the change back to 2018 in src/test/java/net/wildfyre/http/RequestTest.kt and src/test/java/net/wildfyre/users/UserTest.java.

Rebasing on master

Sorry, I can't read any of this. I'm okay with the changes as you shown on Telegram, but I'm not qualified to find bugs

Why is this done server side instead of client side?

Okay. I thought about the other because it's not breaking (and more detailed)

Wait, is this for the Client (web) or app?

Why would we restrict editing?

In T256#4663, @CLOVIS wrote:

This could be replaced by T290, which gets the same results in a non-breaking way.

The lib is NOT ready to be used before 1.0. There's no point trying, there are multiple things that are not ready.

I don't understand what you're trying to do with the resource handling.

Server API already sends a code (known as http status code) errors start with 4 (client errors, e.g. invalid input) or 5 (server errors, e.g. database unavailable)

• Moving User from Java to Kotlin
• Rename .java to .kt
• Moving Users from Java to Kotlin
• Explaining why the let is ?.let and not !!.let & fixing broken Dokka link
• Rename .java to .kt
• Converting LoggedUser to Kotlin
• Added JsonValue.minimalString & JsonValue.prettyString
• Adding LoggedUser.postsList so that Java code can access the posts
• No more parallel unit tests (the testing server can't handle them)
• IssueInTransferException is now able to report on messages that do not parse to JSON
• Better Kotlin integration

In D222#5365, @Hackintosh5 wrote:

Findbugs should be compileonly.