Attribute and CSS-identity helpers for the no-heap XHTML tokenizer: quoted-attribute scanning, <img src> / <a href> capture, CSS element-identity + inline-style extraction, link interning, and embedded-face resolution. More...

#include <stddef.h>
#include <stdint.h>
#include <string.h>
#include "ra8_attributes.h"
#include "ra8_err.h"
#include "reflow.h"
#include "reflow_tokenize_internal.h"

Include dependency graph for reflow_tokenize_attr.c:

Functions
static bool	internal_attr_name_at (const uint8_t tag, size_t i, const char name, size_t name_len)
	True iff attribute `name` (case-insensitive) begins at tag[i].
static bool	internal_attr_quoted_value (const uint8_t tag, size_t tag_len, size_t pos, size_t vstart, size_t *vlen)
	Parse = "value" after an attribute name; return the value span.
bool	priv_reflow_tok_find_attr (const uint8_t tag, size_t tag_len, const char name, size_t name_len, size_t out_voff, size_t out_vlen)
	Locate a named attribute's quoted value span within a tag (no copy).
void	priv_reflow_tok_capture_attr (reflow_t engine, const uint8_t tag, size_t tag_len, const char name, size_t name_len, uint32_t out_off, uint32_t *out_len)
	Copy a named attribute's quoted value from a tag span into the text pool.
ra8_css_element_t	priv_reflow_tok_css_element (reflow_html_tag_t kind, const uint8_t *tag, size_t span)
	Build a CSS element identity from a just-opened tag's attributes.
ra8_css_style_t	priv_reflow_tok_css_inline (const uint8_t *tag, size_t span)
	Parse a just-opened tag's inline style="..." into a CSS declaration.
uint8_t	priv_reflow_tok_intern_link (reflow_t *engine, uint32_t href_off, uint32_t href_len)
	Intern an <a> href slice into the link-target table.
bool	priv_reflow_tok_rel_is_stylesheet (const uint8_t *rel, size_t len)
	True if a rel attribute value contains the stylesheet token.
uint8_t	priv_reflow_tok_resolve_face_slot (const reflow_t engine, const ra8_css_style_t comp)
	Map a cascaded run's (font-family + emphasis) to an embedded face slot.

Detailed Description

Attribute and CSS-identity helpers for the no-heap XHTML tokenizer: quoted-attribute scanning, <img src> / <a href> capture, CSS element-identity + inline-style extraction, link interning, and embedded-face resolution.

Companion translation unit to reflow_tokenize.c. It groups the routines that read a raw <...> start-tag span: locating a named quoted attribute, copying its value into the engine text pool, building the ra8_css_element_t identity (tag + id + class) the cascade matches against, parsing the inline style="..." declaration, interning an <a href> into the link-target table, detecting a <link rel> stylesheet, and mapping a cascaded font-family + emphasis to an embedded face slot. None of these touch the tokenizer's mutable tok_ctx_t; the markup handlers in reflow_tokenize.c call the cross-TU helpers promoted into reflow_tokenize_internal.h. Pure forward-pass scanning – no recursion (NASA P10 Rule 1) and bounded loops (Rule 2).

[Ring 4 / Reflow] {World: NS}

Copyright: Copyright (c) 2026 Brighton Sikarskie SPDX-License-Identifier: MIT

Since: 0.1.0

Definition in file reflow_tokenize_attr.c.

Function Documentation

◆ internal_attr_name_at()

bool internal_attr_name_at	(	const uint8_t *	tag,
		size_t	i,
		const char *	name,
		size_t	name_len )

static

True iff attribute name (case-insensitive) begins at tag[i].

Matches name with a non-name byte (or the leading '<') immediately before it, so a substring like "xsrc" does not match "src" and "xid" does not match "id".

Parameters

[in]	tag	Raw tag span.
[in]	i	Candidate start offset (i + name_len must be in range).
[in]	name	Lower-case attribute name to match.
[in]	name_len	Length of `name`, bytes.

Returns: true iff tag[i..i+name_len) equals name as an attribute name.

Return values

true	Match.
false	No match.

Precondition: tag, name are non-null and tag[i..i+name_len) are readable.; i + name_len <= tag span length.

Postcondition: No state mutated.; Return value depends solely on the inputs.

Note: Pure function; name must already be lower-case.

Since: 0.1.0

Definition at line 66 of file reflow_tokenize_attr.c.

Referenced by priv_reflow_tok_find_attr().

◆ internal_attr_quoted_value()

bool internal_attr_quoted_value	(	const uint8_t *	tag,
		size_t	tag_len,
		size_t	pos,
		size_t *	vstart,
		size_t *	vlen )

static

Parse = "value" after an attribute name; return the value span.

Skips whitespace, requires =, skips whitespace, requires a quote, then scans to the matching quote. The span excludes the quotes.

Parameters

[in]	tag	Raw tag span.
[in]	tag_len	Length of `tag`, bytes.
[in]	pos	Offset just past the attribute name.
[out]	vstart	Receives the value start offset.
[out]	vlen	Receives the value length, bytes.

Returns: true iff a quoted value was found.

Return values

true	vstart / vlen describe the value.
false	No = "..." followed the name.

Precondition: tag, vstart, vlen are non-null.; pos <= tag_len.

Postcondition: On true, [*vstart, *vstart+*vlen) lies within the tag span.; On false, the outputs are unspecified (caller ignores them).

Note: Pure read of tag.

Since: 0.1.0

Definition at line 100 of file reflow_tokenize_attr.c.

References priv_reflow_tok_is_xml_whitespace().

Referenced by priv_reflow_tok_find_attr().

◆ priv_reflow_tok_capture_attr()

void priv_reflow_tok_capture_attr	(	reflow_t *	engine,
		const uint8_t *	tag,
		size_t	tag_len,
		const char *	name,
		size_t	name_len,
		uint32_t *	out_off,
		uint32_t *	out_len )

Copy a named attribute's quoted value from a tag span into the text pool.

Copy a named attribute's quoted value from a tag span into the pool.

Scans the raw <...> span for attribute name (case-insensitive, not preceded by another name character) and copies its quoted value verbatim into the engine text pool. Used for <img src> / <a href> / element id. On no match, a malformed attribute, an empty value, or a full pool the outputs are zeroed, which callers treat as "attribute absent".

Parameters

[in,out]	engine	Engine whose text pool receives the value bytes.
[in]	tag	Raw tag span starting at '<'.
[in]	tag_len	Length of `tag`, bytes.
[in]	name	Lower-case attribute name to find.
[in]	name_len	Length of `name`, bytes.
[out]	out_off	Receives the text-pool offset of the value (0 if none).
[out]	out_len	Receives the value byte length (0 if none).

Returns: Nothing.

Precondition: engine, tag, name, out_off, out_len are non-null.; engine->text_pool_used <= k_reflow_text_pool_bytes.

Postcondition: *out_len > 0 iff a quoted name value was stored in the pool.; The pool grows by *out_len bytes when a value is stored.

Note: Not thread-safe.

Since: 0.1.0

Definition at line 196 of file reflow_tokenize_attr.c.

References k_reflow_text_pool_bytes, memcpy(), priv_reflow_tok_find_attr(), reflow_t::text_pool, and reflow_t::text_pool_used.

Referenced by internal_handle_void(), and internal_open_attrs().

◆ priv_reflow_tok_css_element()

ra8_css_element_t priv_reflow_tok_css_element	(	reflow_html_tag_t	kind,
		const uint8_t *	tag,
		size_t	span )

Build a CSS element identity from a just-opened tag's attributes.

Locates the id and class attribute value spans in place (no copy) so the cascade can match #id / .class selectors; the returned pointers alias tag. A missing attribute leaves its pointer NULL and its length zero. The element's tag enum is stored so type selectors also work.

Parameters

[in]	kind	Classified tag.
[in]	tag	Raw tag span starting at '<'.
[in]	span	Length of `tag`, bytes.

Returns: The element identity for ra8_css_cascade().

Return values

ra8_css_element_t Fully populated struct; absent attributes have NULL pointer and zero length.

Precondition: tag is non-null and holds span bytes.; kind is a valid reflow_html_tag_t value.

Postcondition: No state mutated; returned pointers alias tag.; el.tag == (uint8_t)kind.

Note: Pure read of tag.

Since: 0.1.0

Definition at line 243 of file reflow_tokenize_attr.c.

References ra8_css_element_t::class_len, ra8_css_element_t::class_str, ra8_css_element_t::id, ra8_css_element_t::id_len, priv_reflow_tok_find_attr(), and ra8_css_element_t::tag.

Referenced by internal_handle_start().

◆ priv_reflow_tok_css_inline()

ra8_css_style_t priv_reflow_tok_css_inline	(	const uint8_t *	tag,
		size_t	span )

Parse a just-opened tag's inline style="..." into a CSS declaration.

Locates the style attribute within the raw tag span and delegates to ra8_css_parse_inline() to parse its value into a declaration struct. When the attribute is absent or the parser fails the returned struct has set == 0, indicating no inline declarations override the cascade.

Parameters

[in]	tag	Raw tag span starting at '<'.
[in]	span	Length of `tag`, bytes.

Returns: The inline CSS declaration for this element.

Return values

ra8_css_style_t Struct with set == 0 if no style attribute is present.

Precondition: tag is non-null and holds span bytes.; span > 0 (a zero-span tag cannot carry attributes).

Postcondition: No state mutated.; Returned struct set field is 0 when no style attribute was found.

Note: Pure read of tag.

Since: 0.1.0

Definition at line 279 of file reflow_tokenize_attr.c.

References priv_reflow_tok_find_attr(), and ra8_css_parse_inline().

Referenced by internal_open_styled().

◆ priv_reflow_tok_find_attr()

bool priv_reflow_tok_find_attr	(	const uint8_t *	tag,
		size_t	tag_len,
		const char *	name,
		size_t	name_len,
		size_t *	out_voff,
		size_t *	out_vlen )

Locate a named attribute's quoted value span within a tag (no copy).

The non-copying core of priv_reflow_tok_capture_attr(): returns the value's offset + length *within tag* so callers can either copy it to the pool or scan it in place (e.g. an inline style).

Parameters

[in]	tag	Raw tag span starting at '<'.
[in]	tag_len	Length of `tag`, bytes.
[in]	name	Lower-case attribute name to find.
[in]	name_len	Length of `name`, bytes.
[out]	out_voff	Receives the value offset within `tag`.
[out]	out_vlen	Receives the value length, bytes.

Returns: true iff a quoted name value was found.

Return values

true	out_voff / out_vlen describe the value within `tag`.
false	Attribute absent or malformed.

Precondition: tag, name, out_voff, out_vlen are non-null.; tag_len >= name_len (zero-length searches return false immediately).

Postcondition: No state mutated.; On false, *out_voff and *out_vlen are unspecified.

Note: Pure read of tag.

Since: 0.1.0

Definition at line 153 of file reflow_tokenize_attr.c.

References internal_attr_name_at(), and internal_attr_quoted_value().

Referenced by internal_handle_link(), priv_reflow_tok_capture_attr(), priv_reflow_tok_css_element(), and priv_reflow_tok_css_inline().

◆ priv_reflow_tok_intern_link()

uint8_t priv_reflow_tok_intern_link	(	reflow_t *	engine,
		uint32_t	href_off,
		uint32_t	href_len )

Intern an <a> href slice into the link-target table.

Appends a {href_off, href_len} entry and returns its 1-based id (stored in subsequent text tokens' reserved byte). An empty href or a full table returns 0 (the run renders as plain underlined text – not tappable).

Parameters

[in,out]	engine	Engine whose link-target table grows.
[in]	href_off	Href text-pool offset.
[in]	href_len	Href byte length.

Returns: 1-based link id, or 0 if not interned.

Return values

0	Empty href or the table is full.
1..k_reflow_max_links	The newly assigned 1-based link id.

Precondition: engine is non-null.; engine->link_target_count <= k_reflow_max_links.

Postcondition: On a non-zero return the table grew by one entry.; engine->link_target_count is unchanged on a zero return.

Note: Not thread-safe.

Since: 0.1.0

Definition at line 315 of file reflow_tokenize_attr.c.

References reflow_link_target_t::href_len, reflow_link_target_t::href_off, k_reflow_max_links, reflow_t::link_target_count, and reflow_t::link_targets.

Referenced by internal_open_attrs().

◆ priv_reflow_tok_rel_is_stylesheet()

bool priv_reflow_tok_rel_is_stylesheet	(	const uint8_t *	rel,
		size_t	len )

True if a rel attribute value contains the stylesheet token.

Case-sensitive substring scan (EPUB <link> rels are authored lowercase; this also accepts the alternate stylesheet form). Bounded by the value length. The inner loop iterates at most len times so the scan is O(len) with no dynamic allocation.

Parameters

[in]	rel	Attribute-value bytes (not NUL-terminated).
[in]	len	Length of `rel`, bytes.

Returns: true if stylesheet occurs in rel.

Return values

true	The substring "stylesheet" was found in `rel`.
false	The substring is absent or `len` is too short to contain it.

Precondition: rel addresses len readable bytes.; rel is non-null (a zero len is a valid empty-value case).

Postcondition: No state mutated.; Return value depends solely on the inputs.

Note: Pure.

Since: 0.1.0

Definition at line 347 of file reflow_tokenize_attr.c.

Referenced by internal_handle_link().

◆ priv_reflow_tok_resolve_face_slot()

uint8_t priv_reflow_tok_resolve_face_slot	(	const reflow_t *	engine,
		const ra8_css_style_t *	comp )

Map a cascaded run's (font-family + emphasis) to an embedded face slot.

Returns 0 (the engine's default bound face) when no embedded face is registered, the run has no resolved font-family, no @font-face matches, or the matched @font-face was never registered (its src had no manifest font). Otherwise returns 1 + the registry index, which the render pass uses to pick the matching stbtt_fontinfo.

Parameters

[in]	engine	Engine whose face registry + parsed sheet are consulted.
[in]	comp	The element's cascaded style (family + emphasis bits).

Returns: 0 for the default face, else 1 + engine->faces[] index.

Return values

0	No registered face matches the cascaded family + emphasis combination.

Precondition: engine and comp are non-null.; engine->face_count <= k_reflow_max_faces.

Postcondition: No state is modified.; Return value is in [0, engine->face_count].

Note: Pure; not thread-safe only insofar as it reads engine state.

Since: 0.1.0

Definition at line 382 of file reflow_tokenize_attr.c.

References reflow_t::css, reflow_face_t::css_face_idx, reflow_t::face_count, reflow_t::faces, ra8_css_style_t::family_len, ra8_css_style_t::family_off, k_ra8_css_set_family, k_reflow_style_bold, k_reflow_style_italic, ra8_css_sheet_t::names, ra8_css_match_face(), ra8_css_style_t::set, and ra8_css_style_t::style.

Referenced by internal_open_styled().

Functions

Detailed Description

Function Documentation

◆ internal_attr_name_at()

◆ internal_attr_quoted_value()

◆ priv_reflow_tok_capture_attr()

◆ priv_reflow_tok_css_element()

◆ priv_reflow_tok_css_inline()

◆ priv_reflow_tok_find_attr()

◆ priv_reflow_tok_intern_link()

◆ priv_reflow_tok_rel_is_stylesheet()

◆ priv_reflow_tok_resolve_face_slot()