Properly check if an identical chunk has been written before. #1956
Conversation
src/write.c
Outdated
| chunkOffset = avifEncoderFindExistingChunk(s, mdatStartOffset, sample->data.data, sample->data.size); | ||
| size_t total_size = sample->data.size; | ||
| for (uint32_t sampleIndex = 1; sampleIndex < item->encodeOutput->samples.count; ++sampleIndex) { | ||
| total_size += item->encodeOutput->samples.sample[sampleIndex].data.size; |
There was a problem hiding this comment.
Is the data contiguous?
Suggestion:
AVIF_CHECKERR(
item->encodeOutput->samples.sample[sampleIndex].data.data ==
item->encodeOutput->samples.sample[0].data.data + total_size,
AVIF_RESULT_UNKNOWN_ERROR);(not just an assertion, for extra security)
There was a problem hiding this comment.
Well, the data is written contiguously a few lines below but I'll add a check.
The problem happened starting with #463
There was a problem hiding this comment.
Actually, we do not care right? The stream is already formed so as long as we have a match of size total_size, we are good and the deduplication works.
There was a problem hiding this comment.
I missed it's on the input, that makes sense. Done
Fixes internal bugs: b/321189607 and b/321189607
vrabaud
force-pushed
the
version_number
branch
from
d69c44c to
debabf6
Compare
vrabaud
force-pushed
the
version_number
branch
from
debabf6 to
082c781
Compare
vrabaud
force-pushed
the
version_number
branch
from
a86e26f to
5f791c9
Compare
…aCodec#1956) Basically reverting 295cab3 in AOMediaCodec#463
|
@vrabaud Vincent: The first comment in this pull request says "Fixes internal bugs: b/321189607 and b/321189607", but the two bug numbers are the same. Is there a second bug that this pull request fixes? |
|
Nope, just a mistake on my end ... |
Fixes internal bugs: b/321189607 and b/321189607