Browse the repository at this point in the history

This CL also implements some minimal formatting behavior _around_ `{@snippet ...}` (namely, a blank line before and after). It does not make any effort to change formatting _inside_ `{@snippet ...}`, only to stop reflowing it blindly.

PiperOrigin-RevId: 665555661

Browse the repository at this point in the history

Browse the repository at this point in the history

Browse the repository at this point in the history

Browse the repository at this point in the history

Browse the repository at this point in the history

Browse the repository at this point in the history

* ebnf-mode: Last change 13 years ago with open PR from phst from 2019.  For the sake of the current package version metadata work, just slap in a pkg.el file for now.
* llm-goose: Missing header.
* gyp: Missing header.
* lv: Include pkg.el added earlier.

PiperOrigin-RevId: 667696660

Browse the repository at this point in the history

Browse the repository at this point in the history

Browse the repository at this point in the history

The intent was to only run the test on JDK 23+ and to disable it until the linked issue was fixed, but it actually resulted in the test being discovered and run on all JDK versions.

That's fine, because the test is backwards compatible (`///` comments are valid in all versions, they just aren't markdown comments until 23).

The test is being fixed in #1161

PiperOrigin-RevId: 674352956

Browse the repository at this point in the history

Browse the repository at this point in the history

Despite documentation to the contrary, the column numbers are already 1-indexed. The comment was added in unknown commit when g-j-f was still implemented using ecj instead of javac, so maybe it was true then?

PiperOrigin-RevId: 681451766

Browse the repository at this point in the history

Browse the repository at this point in the history