[nicoburns]

LLM Contributions: Written by Devin (Cognition AI) under direction of @nicoburns.

Rebase of #32 (and the relevant part of #26) onto upstream linebender#790 (hang-conditionally, at cbf2fa1; the base branch here is a verbatim copy of it). Supersedes #32.

Corpus benchmark (250 KB real text, calculate_content_widths, CPU-pinned, interleaved, n=12 x 3):

build	median ms	min ms
pre-linebender#715 (6b92d49, flat ClusterData scan)	0.70	0.68
upstream linebender#790 as-is	1.15–1.31	1.13
this PR	0.77–0.83	0.75

widths-result identical (min: 1160.2813, max: 17458.617). Profile of the new code is a single fully-inlined loop (no Atoms::next / characters[] traffic).

Commit 1 — engine: ShapedClusterFlags has 5 free bits; they now hold the first character's Boundary (2 bits) and Whitespace (3 bits — the current 8-variant enum fits exactly), set in process_shaped_clusters::flush from first_character.info. Additive API on ShapedCluster: boundary_before(), whitespace(), char_len(). ShapedCluster stays 20 bytes. A round-trip unit test covers all variants.

Commit 2 — layout: when the run has no letter/word spacing (the common case) calculate_content_widths scans shaped_clusters directly instead of atoms_start():

• break/newline checks only on atom-start clusters (i == 0 || is_grapheme_start()), with a skip_atom flag to skip a Newline atom's continuation clusters like the atom loop's continue does;
• the hanging accumulator is updated per cluster (if cluster_hangs { += advance } else { = 0 }), which is equivalent to atom_hanging_advance's "hangs entirely / hangs partially from the logical end" rule when there are no gaps;
• whitespace_hangs(first char, style) is checked before char_len() == 1, so letters never load chars_range; multi-char clusters fall back to checking every character (with its own style, as atom_hanging_advance does).
• hangs_conditionally is set from the cluster's style; Implement CSS's conditional hanging of whitespace linebender/parley#790 uses the atom's first-character style, which only differs for mixed-style multi-cluster atoms (the edge case already flagged in Implement CSS's conditional hanging of whitespace linebender/parley#790 review).

Runs with spacing keep linebender#790's atoms_start() loop unchanged (it needs Atom for spacing.atom_advance / atom_hanging_advance).

Note: the first commit's message still mentions the layout change from #32; the layout part is in the second commit.

Changelog: None

Link to Devin session: https://dioxus.staging.devinenterprise.com/sessions/2564f38c1f574d04ac6e9a57a7284d93
Open in Devin Desktop: https://dioxus.staging.devinenterprise.com/desktop/session/2564f38c1f574d04ac6e9a57a7284d93?variant=devin-insiders

[staging-devin-ai-integration]

I'll fix CI failures and address comments from users with write access that start with 'DevinAI' or '@devin'.

• Disable automatic comment, CI, and merge conflict monitoring