Upgrade Sourcemeta dependencies

[jviotti]

Signed-off-by: Juan Cruz Viotti jv@jviotti.com

[cubic-dev-ai]

No issues found across 157 files

<sub>Note: This PR contains a large number of files. cubic only reviews up to 100 files per PR, so some files may not have been reviewed. cubic prioritizes the most important files to review.
On a pro plan you can use ultrareview for larger PRs.

Re-trigger cubic</sub>

[augmentcode]

🤖 Augment PR Summary

Summary: This PR updates vendored Sourcemeta dependencies (Core + Blaze) and incorporates new upstream functionality and canonicalization/compilation improvements.

Changes:

• Bumps pinned revisions for sourcemeta/core and sourcemeta/blaze (plus updated dependency manifests).
• Extends Blaze canonical meta-schemas (notably enum typing constraints) and refactors Draft-3 canonical schema definitions.
• Adds new Blaze canonicalizer rules (disallow_double_negation, enum_split_by_type) and wires them into the alterschema pipeline.
• Adds compiler support for selectively emitting annotations in exhaustive mode via Tweaks::annotations and centralizes the gating logic.
• Improves Blaze schema walking/framing to detect embedded meta-schemas in nested resources.
• Upgrades Core with new text utilities (ASCII is_alpha/is_digit/is_alphanum) and introduces a new langtag component for BCP 47 validation.
• Enhances Core URI handling with IRI support (parse/recompose/canonicalize), adds utf8_decode, and updates URI equality/ordering semantics.
• Adds URITemplateRouter::describes / view equivalent to query whether a path is covered by registered routes.
• Improves thread-safety of Markdown rendering by serializing cmark-gfm parser construction/teardown.

Technical Notes: Most code changes are vendored upstream improvements (explicit lambda return types, designated initializers, and safer string-view handling) with a few notable behavioral additions around IRIs, routing introspection, and annotation emission control.

_{🤖 Was this summary useful? React with 👍 or 👎}

[augmentcode]

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

[augmentcode]

vendor/core/src/core/yaml/yaml.cc:44 — Catching YAMLParseError and rethrowing YAMLFileParseError will erase more-specific derived exceptions (e.g. ones carrying extra context like an anchor name), potentially reducing diagnosability. Consider whether you want to preserve specialized error details while still attaching the path.

Severity: medium

Other Locations

• vendor/core/src/core/yaml/yaml.cc:81

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}

[augmentcode]

vendor/core/src/core/uri/path.cc:35 — canonicalize_path now accepts literal non-ASCII (RFC 3987) bytes, but it still only percent-decodes RFC 3986 unreserved characters, so semantically equivalent IRI paths like /caf%C3%A9 vs /café may not compare the same in strip_path_prefix/rebase_path. Consider whether these helpers should normalize percent-encoded UTF-8 iunreserved sequences consistently when Unicode is allowed.

Severity: medium

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}