-
Notifications
You must be signed in to change notification settings - Fork 29k
[SPARK-30164][Tests][Doc] SBT clean before document generation #26796
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
Conversation
|
Test build #114985 has finished for PR 26796 at commit
|
| # Enable all of the profiles for the build: | ||
| build_profiles = extra_profiles + modules.root.build_profile_flags | ||
| sbt_goals = ["unidoc"] | ||
| sbt_goals = ["clean", "unidoc"] |
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.
Hm.. I avoided this because it will slow down doc gen since it re-uses complied ones from the regular build. How long does it increase?
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.
Also, if I am not wrong, this wont verify the documentation in test codes anymore. That's fine as documentation in the test codes wont be the part of our official documentation though. Cc @JoshRosen I think I talked about this with you before somewhere.
docs/README.md
Outdated
| ## API Docs (Scaladoc, Javadoc, Sphinx, roxygen2, MkDocs) | ||
|
|
||
| You can build just the Spark scaladoc and javadoc by running `./build/sbt unidoc` from the `$SPARK_HOME` directory. | ||
| You can build just the Spark scaladoc and javadoc by running `./build/sbt clean unidoc` from the `$SPARK_HOME` directory. |
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 can imagine adding this to the docs, as users might have stale compiled code locally or something. How often would it occur in run-tests? In PR builders it starts with a clean build and I'd guess users recompile before tests. I'm also worried about slowdown for the PR builders.
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.
@srowen Thanks for the suggestion. I will revert the changes here.
|
@HyukjinKwon @srowen Yes we should see if there is another approach that can avoid significantly increasing the PR builder execution time With the code change: 233s (https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/114990) |
|
Currently, this blocks #26788 . |
|
Does this actually affect Jenkins? it starts from a clean build, I'd imagine. I thought this might simply affect end users. |
|
I think it will re-compile again when we run the actual tests ...
Yeah, seems like it. We might have to reorder unidoc .. but I wonder if we have a better option. Let me take a cursory look today if this isn't urgent .. |
|
Thank you, @HyukjinKwon ~ |
I would like to try that, too. But actually I don't know the reason for build failure. The two actions "sbt assembly/package" and "sbt unicode" use the same profiles... Anyway, @HyukjinKwon I have tried the reorder in the other PR: 970ec48 |
|
Thanks. Actually, I am trying a different way at #26800 too .. let me see if the tests pass. Hive domain isn't our official documentation anyway so we could simply just exclude. (see also https://github.com/apache/spark/blob/master/docs/_plugins/copy_api_dirs.rb#L30) |
|
Test build #114998 has finished for PR 26796 at commit
|
|
I am closing this one, in favor of #26800 |
What changes were proposed in this pull request?
In the latest master branch, the document generation command:
fails with the folllowing message:
To fix it, we should change "sbt unidoc" to "sbt clean unidoc"
This PR also updates related README.
Why are the changes needed?
Fix document generation in testing
Does this PR introduce any user-facing change?
No
How was this patch tested?
Manually test locally.
Also in #26788, the Jenkins tests pass with the fix.