tree 7620031cebde57b95c115270bf225a5b644ad680
parent e2c30849a21d0f849e7196402c267d66551d48a3
author Miklos Vajna <vmiklos@collabora.co.uk> 1607535773 +0000
committer Chromium commit bot <commit-bot@chromium.org> 1607535773 +0000

FPDF_GetTrailerEnds: make this not depend on whitespace

PDF-1.7 calls out no bytes other than whitespace when specifying what
can occur between endstream and endobj, so whitespace needs to be
handled. When CPDF_SyntaxParser::ReadStream() reads the stream, it reads
'endobj', and then resets the position back to the end of 'endstream'.
This mechanism is disabled in case there is whitespace between the
tokens and the newline, see the end of the function.

This results in reporting no trailer ends, as the parsing fails, as the
next token is expected to be 'endobj', but it's the ID of the next
object instead.

Fix the problem by handling whitespace in
CPDF_SyntaxParser::ReadStream() where it was looking for \ntoken\n, not
allowing whitespace between the token and the following newline.

Change-Id: I7048e8d081af04af3dd08d957212c885b7982b5e
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/76850
Commit-Queue: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>