[SPARK-20027][DOCS] Compilation fix in java docs.

[ScrapCodes]

What changes were proposed in this pull request?

During build/sbt publish-local, build breaks due to javadocs errors. This patch fixes those errors.

How was this patch tested?

Tested by running the sbt build.

[SparkQA]

Test build #74863 has started for PR 17358 at commit 4351840.

[HyukjinKwon]

Hi @ScrapCodes, I just happened to look at this PR and ran jeykill build against this PR.

It seems there are few more instances to make javadoc8 working as below:

[error] .../spark/mllib/target/java/org/apache/spark/ml/fpm/FPGrowthParams.java:17: error: bad use of '>'
[error]    * Number of partitions (>=1) used by parallel FP-growth. By default the param is not set, and
[error]                            ^
[error] .../spark/sql/core/src/main/java/org/apache/spark/api/java/function/FlatMapGroupsWithStateFunction.java:30: error: reference not found
[error]  * {@link org.apache.spark.sql.KeyValueGroupedDataset#flatMapGroupsWithState(
[error]           ^
[error] .../spark/core/target/java/org/apache/spark/TaskContextImpl.java:10: error: reference not found
[error]  * {@link TaskMetrics} & {@link MetricsSystem} objects are not thread safe.
[error]           ^
[error] .../spark/core/target/java/org/apache/spark/TaskContextImpl.java:10: error: reference not found
[error]  * {@link TaskMetrics} & {@link MetricsSystem} objects are not thread safe.
[error]

Should we maybe sweep these together if I found them correctly?

[srowen]

This should just be a sentence rather than fragment.

[ScrapCodes]

Thanks !

[ScrapCodes]

Hi @HyukjinKwon !, Thank for looking at this. I did one pass through all those errors, they all seem to be inside the generated code. Please notice /target in the path.

[HyukjinKwon]

Yea, It generates java codes for Java API. I suggested to fix these together because they look similar instances and I didn't want to open another minor PR for javadoc8 (BTW, FlatMapGroupsWithStateFunction.java seems not a generated code).

If you think this should be separately handled in another PR, then I am happy to do so.

If you meant it is not an issue,

I tried both

 rm -fr target
 ./build/sbt clean
 ./build/sbt unidoc

and

cd docs && jekyll build

I see

[error] (spark/javaunidoc:doc) javadoc returned nonzero exit code

and

jekyll 3.3.1 | Error:  Unidoc generation failed

due to those four errors and I don't see _site and the Java API documentation after jekyll build as written in docs#generating-the-documentation-html.

[SparkQA]

Test build #74883 has finished for PR 17358 at commit ab3fb76.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

If my suggestion is the reason to hold off this PR currently and you are not sure of it, I am fine. I can double-check and deal with this in a separate PR.

[ScrapCodes]

@HyukjinKwon, Fixing those errors require fixing unidoc plugin's java code generation. IMO, that has a broader scope. What do you think ?

[SparkQA]

Test build #75028 has started for PR 17358 at commit 1338a75.

[HyukjinKwon]

Sorry for making this PR dragged. I am fine. Let me cope with it separately.

[ScrapCodes]

Actually thank you for your comments, I missed one file in swarm of errors from generated files.

[ScrapCodes]

Jenkins, retest this please.

[SparkQA]

Test build #75035 has finished for PR 17358 at commit 1338a75.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #3605 has finished for PR 17358 at commit 1338a75.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[srowen]

Merged to master