Issue #14573: fix javadoc type check for empty tags

[strkkk]

Closes #14573

Diff Regression config: https://gist.githubusercontent.com/strkkk/26466254f89c32d8f1a6bd53b6f5b251/raw/cf451b3aa6366addc027a08423786d455bc79aa7/config_single.xml

[strkkk]

Github, generate report

[github-actions]

https://checkstyle-diff-reports.s3.us-east-2.amazonaws.com/e5f329e_2024235419/reports/diff/index.html

[strkkk]

There are cases like this https://checkstyle-diff-reports.s3.us-east-2.amazonaws.com/e5f329e_2024235419/reports/diff/spotbugs/index.html#A1, when annotation is considered as tag. In theory, we can analyse context and ignore text inside <code> or <pre> tags

[strkkk]

Problem is in BlockTagUtil#extractBlockTags, it does not take context into consideration. There are 2 possible solutions:

1. Consider tags

    and  and ignore its content

2. Adjust regex to make it parse tag only if it starts with lowercase letter. I couldn't find some article / spec to back it up, but it seems reasonable

[romani]

This Check is not AST based, so amount of problems in this Check is uncountable and we will never fix them. We need to migrate to AST.

Without migration, we just need to make sure that diff report is decently ok or better then before in behavior

[strkkk]

Without migration, we just need to make sure that diff report is decently ok or better then before in behavior

it seems ok to me if we ignore cases with code snippets with annotations in javadocs

[romani]

openjdk17 ( ±1954, +1954 )

this is too much, we can NOT implSpec, apiNote to known tags in this Check in config.
Please add filter to you config_single.xml to skip such violations, and get cleaner report.
We will need to post this filter in migration notes, as such tags become more and more popular, we already have issue somewhere to add them to support.

Consider tags and and ignore its content

How complicated to add this ? User do not heavily like false positives, and "ok" with false-negatives, especially in javadoc.

[romani]

This is where extra property is requested #14724 (comment) it will ease a problem, but it is separate issue.