Video controls are now hidden by default when playback begins.
Users can tap to reveal them. Previously, controls were shown for
2 seconds before auto-hiding (TwoSecondController) or shown by
default (VideoViewInner).
https://claude.ai/code/session_01WpAqbodB5w8oDJpN6MLBs6
MetadataStripper, MediaCompressor, and EncryptFiles create intermediate
temp files in cacheDir during the upload pipeline, but these were never
deleted after the upload finished. Over time this causes unbounded disk
growth. Wrap upload calls in try/finally to ensure all intermediate temp
files (compressed, stripped, encrypted) are deleted regardless of
success or failure.
https://claude.ai/code/session_01YS3dbZ1k84N2EHS3AruYRV
The progress scrubber can now be dragged to seek through the video,
in addition to the existing tap-to-seek. During drag, the scrubber
enlarges for visual feedback and tracks the finger position in
real-time. Seeking is applied on drag end.
https://claude.ai/code/session_01QvYAvNdC35zGgkdpnujFn2
ChaCha20Core.chaCha20Xor:
- Parse key/nonce once into initial state, reuse across blocks
- XOR at word level (4 bytes at a time) instead of byte-by-byte
- Reuse working IntArray across blocks instead of allocating per block
XChaCha20Poly1305:
- Add hChaCha20FromNonce24 to avoid nonce copyOfRange(0,16) allocation
- Add chaCha20PolyKey to generate only 32 bytes instead of full 64-byte block
- Use single result array instead of ciphertext + tag concatenation
Benchmark (98-byte padded message, JVM):
Before: encrypt 2286 ns/op, decrypt 2062 ns/op
After: encrypt 1544 ns/op, decrypt 341 ns/op
https://claude.ai/code/session_01YHBwqnb3Z7Q7PX31tJ2G3i
Tests the StringIndexOutOfBoundsException fix in Url.getPart() by directly
constructing Url objects with UrlMarker indices that exceed the trimmed
originalUrl length — the exact scenario that occurs when readEnd() strips
a trailing character (e.g. ':') while the PORT/QUERY marker still points
to the original buffer position.
https://claude.ai/code/session_013rJ9iYndYVJgpYK2VazxL2
Tests cover the StringIndexOutOfBoundsException in Url.getPart() that
occurred when readEnd() trimmed trailing characters (e.g. ':') from a
detected URL but urlMarker indices remained pointing past the trimmed
string's end.
- UrlMarkerTest: three cases testing PORT/QUERY index at or beyond
string length (the boundary conditions fixed by minOf clamping and
the startIndex >= length guard)
- UrlsDetectorTest: regression for "今北産業" (no crash, 0 URLs) and
for a bare "example.com:" host
- UrlParserTest: end-to-end regression for "今北産業" and "example.com:"
https://claude.ai/code/session_013rJ9iYndYVJgpYK2VazxL2
When UrlDetector.readEnd() trims the last character from URLs ending
with characters like '.', ',', etc., the urlMarker indices remain
based on the original buffer length. This caused getPart() to call
substring() with an end index beyond the trimmed URL's length.
Fix by clamping the end index to originalUrl.length and guarding
against an out-of-bounds start index.
https://claude.ai/code/session_014Q2SvTE5vKgUm1w8dPVjYo
* 'main' of https://github.com/vitorpamplona/amethyst:
New Crowdin translations by GitHub Action
update translations: CZ, DE, PT, SE
New Crowdin translations by GitHub Action
JesterContent fields version, fen, and history were omitted from
serialized JSON because kotlinx.serialization skips default values.
Jester's isStartGameEvent checks arrayEquals(json.history, []) which
fails when the field is absent (undefined !== []).
Added @EncodeDefault to version, fen, and history so they are always
included in the JSON output regardless of value.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Claude