tree 0b02e78403554e9e90db3ac13199dfe9aa538fd1
parent b62e4269829cfe3b019c3c494227409d72cc4eb4
author Lei Zhang <thestig@chromium.org> 1746747682 -0700
committer Pdfium LUCI CQ <pdfium-scoped@luci-project-accounts.iam.gserviceaccount.com> 1746747682 -0700

M137: Revert "Rework inline image parsing in content streams"

This reverts commit c0bc98d0cacef2a77767aeee4a549a97271b9c4f.

Reason for revert: Broke inline image parsing for other PDFs

Bug: 407752631, 414588524, 416278203
Original change's description:
> Rework inline image parsing in content streams
>
> Currently, CPDF_StreamContentParser:Handle_BeginImage() will try to
> parse inline images in content streams when it sees the "BI" operator.
> The parts that deal with the related "ID" and "EI" operators
> (Handle_BeginImageData() and Handle_EndImage(), respectively) are
> no-ops, since Handle_BeginImage() did all the work.
>
> Rework Handle_BeginImage() and Handle_BeginImageData() so
> CPDF_StreamContentParser records their respective keywords's positions.
> Then handle the inline image parsing in Handle_EndImage() with those
> positions. At this point, CPDF_StreamContentParser know the distance
> between the "ID" and "EI" operators, so it knows where the length of the
> inline image stream. Pass this length to
> CPDF_StreamParser::ReadInlineStream() and prevent it from reading the
> "EI" operator or past it.
>
> As a result of this change:
> - Handle_BeginImage() no longer needs to look for "ID" and "EI".
> - ReadInlineStream() no longer needs to looks for "EI".
> - bug_1236805.in fails because the code was reading "EI" and past it.
>   Fix this by adding a placeholder value so the inline stream is not
>   empty. Remove the temporary suppression for this file.
> - The bug_407752631.in test case can be checked in and will no longer
>   cause a crash.
>
> To deal with malforms images that may not have "BI", "ID", and "EI"
> operators in that order, do some state tracking so that
> CPDF_StreamContentParser only looks for "ID" and "EI" when appropriate,
> like the existing code.
>
> Bug: 407752631
> Change-Id: Ie14908d8ed72c92c2ae880c18f62469a5cf4eef3
> Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/130911
> Commit-Queue: Lei Zhang <thestig@chromium.org>
> Reviewed-by: Tom Sepez <tsepez@chromium.org>

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 407752631
Change-Id: I4860c988fc0cb0fc5b1add9c57b0806378b1e1e6
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/131770
Commit-Queue: Lei Zhang <thestig@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
(cherry picked from commit 4d936fae88e88fccdc1422bab23d5b10022a4036)
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/131950