build: github-CI is currently broken due to missing plpgsql build rules

[knz]

Describe the problem

The "Github CI" / "Lint" target in TC is currently failing on master with the following error:

[15:47:57] :			 [run] # github.com/cockroachdb/cockroach/pkg/sql/scanner
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:35:23: undefined: lexbase.SCONST
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:43:23: undefined: lexbase.IDENT
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:50:23: undefined: lexbase.SCONST
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:60:24: undefined: lexbase.BCONST
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:71:23: undefined: lexbase.DOT_DOT
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:83:23: undefined: lexbase.NOT_EQUALS
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:92:23: undefined: lexbase.LESS_LESS
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:96:23: undefined: lexbase.LESS_EQUALS
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:105:23: undefined: lexbase.GREATER_GREATER
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:109:23: undefined: lexbase.GREATER_EQUALS
[15:47:57] :			 [run] pkg/sql/scanner/plpgsql_scan.go:109:23: too many errors

To Reproduce

Example here: https://teamcity.cockroachdb.com/buildConfiguration/Cockroach_UnitTests_Check/9610026?showRootCauses=false&expandBuildChangesSection=true&expandBuildProblemsSection=true

Expected behavior

No error

Jira issue: CRDB-27002

[e-mbrown]

Weird, this may be caused by go code not being generated before this linter runs. I'll take a look today.

[rickystewart]

Regressor seems to be 19b6cb96436a5cc7a5c25632f15b4b5df496d6bd.

[rickystewart]

The commit seemed to introduce some non-trivial build logic so we would have to port it over to make in this case.

[rafiss]

Are we still supporting the Makefile on master? If not, when would we stop running the Github CI job?

[knz]

The linter in github CI is still catching more than the one in bazel. we have a bazel item left to investigate that, it's AFAIK the remaining blocker.

[rickystewart]

GitHub CI still does GCAssert. #65485

Unfortunately I tried to set up a Bazel-based GCAssert job but it was flaky and no one has been able to identify the source of the flakiness.