Pinned parser rejects EO examples causing Incorrect examples to not trigger lint

What happens

Many EO examples in src/main/resources/org/eolang/motives/**/*.md are not valid EO for the parser that lints depends on (eo-parser 0.63.0, home.version in pom.xml). In 22 motives the Incorrect example doesn't even trigger the lint it documents, because the parser rejects it before the lint gets a chance to run.

I ran every fenced eo and xml block of every motive through EoSyntax(...).parsed() and then through its own LtByXsl (lints de102d68). There are 199 blocks, and 85 of them, in 41 motives, come back with parser errors:

Most of them are the old style of putting a # Comment. right above an object, QQ.io.stdout, and tests written as [] +> name without the blank line.

The Incorrect example yields no defect of its own lint in these motives:

atoms/atom-without-rt, comments/comment-is-too-wide, comments/comment-not-capitalized, comments/comment-starts-with-article, comments/comment-too-short, comments/comment-without-dot, critical/named-object-abstract-nested, design/no-attribute-formation, design/too-deep-object, misc/anemic-getter, misc/incorrect-test-object-name, misc/redundant-object, misc/sparse-seq, misc/sprintf-constant-args, misc/sprintf-without-formatters, misc/wrong-sprintf-arguments, names/anonymous-formation, names/compound-name, names/invalid-name-notation, names/phi-is-not-first, tests/test-word, tests/unit-test-without-phi.

In atoms/rt-without-atoms, metas/incorrect-jvm-rt-location, metas/incorrect-node-rt-location and tests/unit-test-missing, it's the other way round: the Correct example doesn't parse, and its own lint reports a defect on it.

For example, comments/comment-too-short.md:

# This comment is short.
[] > foo
  42 > @

is rejected with [2:0] a blank line must separate the top comment block from the rest of the file, and comment-too-short stays silent on it.

Why it matters

The motive is what a user reads after getting a defect (Lint.motive()). An example that the compiler refuses, or one that the lint doesn't flag, doesn't explain the rule. It also hides real drift. For instance, the three sprintf motives still use QQ.txt.sprintf, which is the base #1371 says the modern parser never emits.

Nothing tests the motives. LtByXslTest and PkMonoTest only check that the file exists and isn't empty.

How to fix

1. Rewrite the examples in current EO syntax: comments only in the file header followed by a blank line, Q.org.eolang... or plain names instead of QQ, tests as +> attributes with exactly one blank line before them.
2. Add a test that walks every motive, parses each block under Incorrect: and Correct:, fails on parser errors, and checks that the lint reports at least one defect on the Incorrect blocks and none on the Correct ones. That's the check above. It took about 20 seconds for all 94 motives.