tballison commented on PR #2947:
URL: https://github.com/apache/tika/pull/2947#issuecomment-4959315401
Agent has one small nit:
```
TikaUserDataBox.java:164-167 — the disk branch reads a fixed 6 bytes with no
length guard, unlike the adjacent
trkn branch (PLAUSIBLE, pre-existing but in the touched branch). The trkn
case guards if (toRead == 8) before
reading 8 bytes and skips otherwise; the disk case unconditionally reads
getInt32()+getInt16() (6 bytes) then
does totalRead += toRead. If any encoder writes a disk atom whose data
length ≠ 6 (e.g. an 8-byte form mirroring
trkn), totalRead diverges from the bytes actually consumed and the
remaining ilst records are read misaligned.
Real-world files are 6 bytes so this is latent, but since the PR now reads
b from this branch it's the natural
spot to add the same toRead ==-style guard/skip the trkn branch already
has.
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
