Fix the PGXN tag that blocked the 2.0.0 upload, and check META.json in CI

[jdatcmd]

PGXN refused the 2.0.0 distribution:

Field /tags[11]: "pl/sql" is not a valid tag [Spec v1.0.0]

A Tag may not contain a slash. Now plsql. No other tag was at fault.

Versioning

The distribution version moves to 2.0.1; the extension version stays at 2.0.0, because nothing about the extension changed. So there is no plx--2.0.1.sql and no upgrade script — adding a version-advancing script and a duplicate install script for a metadata typo would leave dead files in the catalog forever. PGXN treats the two versions as separate, which is exactly this case.

There is no ALTER EXTENSION plx UPDATE for this release, and the CHANGELOG says so.

The actual problem

The only thing validating META.json was PGXN, at upload time — after a release is tagged and published. That's an expensive place to find a typo.

test/check_meta.py (make metacheck) now checks the spec's constraints and runs in CI ahead of the regression suite. Against the unfixed file it reproduces PGXN's rejection exactly:

META.json does not satisfy the PGXN Meta Spec v1:
  tags[11] 'pl/sql' must not contain a slash
exit=1

Why this was easy to get wrong

Worth recording, because the prose and the schema disagree and the prose is the trap.

The spec text says a Tag may contain no "slash, backslash, control, or space" characters. That sentence describes a Term. The Tag schema is:

^[^/\\\p{Cntrl}]{2,}$

which permits spaces. So sql server and string builder are valid tags that would be invalid terms, and the slash in pl/sql was the only real fault. Reading the prose would have led to "fixing" three tags that were never broken.

The validator implements the schema rather than the prose, and its docstring says why, so the next person doesn't repeat that.

Verification

PG18.4: metacheck passes, 13/13 installcheck. Every tag re-checked against the real pattern; 15 tags, all valid.

🤖 Generated with Claude Code