fix(ebcdic): require record names to be unique

[aniketwaghh]

EbcdicLayout requires record selectors to be unique but says nothing about record names, and
EbcdicRecordLayout.name defaults to "record". A copybook with two schemas, unique selectors and
no names is therefore accepted, and _RecordParser.parse buckets the decoded rows by name:

rows = {record.name: [] for record in self._layout.records}

Records are selected by selector but stored by name, so both schemas share one bucket. convert()
then hands each schema that combined list while _build_table writes the schema's own header over
it, and every table ends up holding every record. The conversion reports ConversionStatus.SUCCESS.

This adds the missing check to _validate_records, next to the selector one it mirrors. It also
resolves the two identical ## record headings such a layout produced, since name is what
add_heading prints.

Before, on main at 97bdf30, converting an 84-byte cp037 file holding two customer records tagged
C and two transaction records tagged T:

## record

| customer_name    |   account_no |
|------------------|--------------|
| ACME SUPPLIES    |         0041 |
| BOREALIS TRADING |         0042 |
| INVOICE 88213    |      1250.50 |
| REFUND  88240    |       -49.99 |

## record

| memo             |   amount |
|------------------|----------|
| ACME SUPPLIES    |     0041 |
| BOREALIS TRADING |     0042 |
| INVOICE 88213    |  1250.50 |
| REFUND  88240    |   -49.99 |

A transaction amount of 1250.50 is presented as a customer's account_no, and the account number
0041 as a monetary amount. After, the same layout raises at construction:

pydantic_core._pydantic_core.ValidationError: 1 validation error for EbcdicLayout
  Value error, record names must be unique

Adding name="customer" and name="transaction", changing nothing else, converts to the two
correct tables both before and after this change.

I considered keying rows by the schema's index instead, which would fix the data without rejecting
any layout. I went with the validator because name is also the table heading, so two schemas
sharing one is ambiguous output even when the rows are right — but say the word if you would rather
not reject layouts that are currently accepted, and I will switch it.

Three tests. test_record_names_must_be_unique and test_record_names_must_be_unique_from_a_layout_file
cover the two ways a copybook arrives, inline and as JSON; both fail on main and pass here.
test_a_single_record_keeps_the_default_name guards the case the check must not break — one schema
keeping the "record" default — and passes either way, so it is a regression guard rather than a
bug test.

Suite on this branch and on main at 97bdf30, same machine: 12 failed / 1777 passed here against
12 failed / 1774 passed there, the three extra passes being the tests above. The twelve failures are
the same twelve on both sides — asr_pipeline, webp, the two e2e conversions and the two extraction
template tests, all model- or network-dependent. make validate is clean.

Issue resolved by this Pull Request:
Resolves #4125

Checklist:

• Documentation has been updated, if necessary.
• Examples have been added, if necessary.
• Tests have been added, if necessary.

[github-actions]

✅ DCO Check Passed

Thanks @aniketwaghh, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🟢 Merge protection satisfied — ready to merge.

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?: