NodetteHQ/PureYAML
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

docs: record the libyaml port as complete and the sole parsing engine

Browse Source 
Marks every step of the port plan done and adds a status section: the port
is integrated as PureYAML's only parser, verified byte-for-byte against
libyaml on 307/308 cases, and scores 275/308 on the canonical event gate,
ahead of libyaml's own 252 and the prior hand-rolled parser's 267.
This commit is contained in:
Mihaela Mihaljevic 2026-06-24 18:17:12 +02:00
parent 18101b8ddc
commit 884ff7d9a1
1 changed files with 31 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
docs/LIBYAML_PORT.md
View File

@ -1,5 +1,28 @@
# Faithful libyaml port
## Status: complete and integrated
The port is finished and is now PureYAML's **only** parsing engine. Every step
below has landed, the hand-rolled approximation has been deleted, and `parseEvents`,
`parseStreaming`, and `parseStreamStreaming` all run on the port.
- **Faithfulness**: verified byte-for-byte against the libyaml reference event
dumper across all 308 vendored valid cases. The port matches libyaml on 307; the
single difference (`L24T/01`) is the port being more correct than libyaml.
- **Conformance** (canonical `test.event` gate, of 308): libyaml itself scores
252, the old hand-rolled parser scored 267, and the port scores **275** after
closing real YAML 1.2 gaps libyaml lacks (reserved directives ignored with a
warning, newer `%YAML` minor versions, multi-line flow keys, bare documents
after `...`, tab-only blank lines, JSON UTF-16 surrogate-pair `\u` escapes, full
`ns-anchor-char` anchor/alias names).
- **Integration boundary** (`Sources/Parsing/Parser.Events.swift`): converts
libyaml's 0-based marks to PureYAML's 1-based, translates scanner/parser errors
to the diagnostic-aware `ParseError`, and enforces a 64-level nesting bound so
deep input throws `nestingTooDeep` rather than overflowing the recursive composer.
Remaining open gaps and their clusters are tracked by `knownEventGaps` in
`Tests/CanonicalEventConformanceTests.swift`.
## Why
PureYAML's original scanner is a hand-rolled approximation of libyaml's
@ -33,17 +56,16 @@ is a translation of the algorithm into pure Swift, not a binding to libyaml.
more indented than the current level, which fixes the `- a:` / same-column `b:`
bug by construction. Implemented in `Sources/Parsing/LibYAML.IndentEngine.swift`,
unit-tested in `Tests/LibYAMLIndentEngineTests.swift`.
2. **Simple-key mechanism**: `save_simple_key`, `remove_simple_key`,
2. **Simple-key mechanism** (DONE): `save_simple_key`, `remove_simple_key`,
`stale_simple_keys`, the `simple_keys` stack and `simple_key_allowed` flag.
3. **Token-fetch core**: `fetch_more_tokens`, `fetch_next_token`,
3. **Token-fetch core** (DONE): `fetch_more_tokens`, `fetch_next_token`,
`scan_to_next_token`, and the `fetch_*` dispatch (`fetch_key`, `fetch_value`
with its empty-scalar and roll-indent logic, `fetch_block_entry`,
`fetch_flow_*`, document indicators, stream start/end).
4. **Scalar scanners**: `scan_plain_scalar`, `scan_flow_scalar`,
4. **Scalar scanners** (DONE): `scan_plain_scalar`, `scan_flow_scalar`,
`scan_block_scalar` (+ breaks), `scan_anchor`, `scan_tag`, `scan_directive`.
5. **Reader**: confirm the existing `Reader` covers `CACHE`/`SKIP`/`READ` and
UTF-8 marks, or port the needed pieces.
6. **Parser**: port `parser.c` to compose events from the token stream.
7. **Integration**: route `parseEvents` / `parseStream` through the port; delete
the approximation once the canonical event gate and libyaml differential are
green with zero regressions.
5. **Reader** (DONE): scalar-based `Reader` with `peek`/`skip`/`skipLine` and marks.
6. **Parser** (DONE): `parser.c` ported to compose events from the token stream.
7. **Integration** (DONE): `parseEvents` / `parseStream` run on the port and the
approximation has been deleted; the canonical event gate and libyaml
differential are green with zero regressions.