The temporal direct derivation truncated (mvCol · DistScaleFactor + 128) >> 8
with `as i16`, wrapping mvL0 past the spec's motion-vector range on crafted
streams, and mvL0 − mvCol then panicked a debug build with "attempt to
subtract with overflow". The scale and the subtraction now run in i32 (dsf is
already clamped to ±1024, so they cannot overflow) and only the results are
clipped to −2^15..2^15−1, as ITU-T H.264 8.4.1.2.4 reads — mvL1 scaled from
the unclamped mvL0, the shape review settled for the same decoder code in
SoundCraft (soundcraft#994): dsf 512, mvCol 32767 keeps mvL1 at 32767.
Extends the grouped robustness audit's codec-overflow list (#1183), which
covers the POC/MMCO/latency sites but not this one.
Co-Authored-By: GLM-5.3 <noreply@z.ai>
The dequantisation wraps by design (huge_levels_dequantise_without_
overflow_panics), so idct4_add and idct8_add receive any i32 coefficient,
and their plain-i32 butterflies panicked a debug build with "attempt to
add with overflow" on extreme values. A dequantised r outside
-2^15..2^15-1 leaves the transforms' output unspecified (ITU-T H.264
8.5.12.1, 8.5.13.1), so those inputs are clamped to that range —
bit-exact for every conforming stream — instead of panicking.
Debug-only (release wraps), memory-safe either way; the sibling SoundCraft
decoder fixed the same class (its audit's H.264 arithmetic findings).
Extends the grouped robustness audit's codec-overflow list (#1183), which
covers the POC/MMCO/latency sites but not these transforms.
Co-Authored-By: GLM-5.3 <noreply@z.ai>
Grouping a group again and again had no limit, but a .deckcraft saved
with 39 or more nested groups could not be opened again (the JSON
reader's recursion limit), and past 64 levels shape walks such as
format.languageAll stop. arrange.group now refuses to nest past 32,
the depth the PPTX reader reads and the renderer draws, and leaves the
slide unchanged. The test builds the deepest allowed deck (a table at
the bottom) and checks it saves and reopens.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
In a narrow window a long language name ("… (not checked)") painted
over the Notes and Comments buttons, and its click area covered them.
The label is now clipped where the controls begin, and only its visible
part opens Review › Language.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The transient message ("Saved") was painted at a fixed x of 290, so
"English (United States) (not checked)" ran under it. The language and
the message now each start at their usual column or after the item
before them, and the message is clipped before the controls on the
right.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An unclamped hostile selection end (text.select caret [1e18, 0]) made
covered_paragraphs iterate for effectively forever.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Otherwise presenter_grid stayed true and held auto-advance and rehearsal for the
rest of the show (CodeRabbit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A Wingdings or Webdings face without private-use coverage no longer gets
Symbol's mapping (0xD8 drew as ¬).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The format repair clamps a crop to 0..1, but an unrepaired negative value still
took the stretch path in paint_frame (CodeRabbit on #1008).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>