[gs-bugs] [Bug 698785] - MuPDF - Can't open this pdf document

[b***@artifex.com]

http://bugs.ghostscript.com/show_bug.cgi?id=698785

Bug ID: 698785
Summary: Can't open this pdf document
Product: MuPDF
Version: 1.11
Hardware: PC
OS: Linux
Status: UNCONFIRMED
Severity: normal
Priority: P4
Component: mupdf
Assignee: mupdf-***@artifex.com
Reporter: ***@abv.bg
QA Contact: gs-***@ghostscript.com
Word Size: ---

Created attachment 14497
--> http://bugs.ghostscript.com/attachment.cgi?id=14497&action=edit
Contains text in Bulgarian and mathematics formulae

The attached file can't be opened with MuPDF but other pdf reading software
such as Kde's okular opens it just fine without any errors.

While trying to open the file with MuPDF a window opens with the following
message on it:
warning: Errors found on page

The following messages are also printed on the terminal:
error: unknown keyword: 'pagesize'
error: unknown keyword: 'width'
error: unknown keyword: 'height'
mupdf: warning: Errors found on page

This happens atleast on versions 1.10 and 1.11

[b***@artifex.com]

http://bugs.ghostscript.com/show_bug.cgi?id=698785

--- Comment #1 from Veselin Ivanov <***@abv.bg> ---
Created attachment 14498
--> http://bugs.ghostscript.com/attachment.cgi?id=14498&action=edit
This is the document rendered in KDE's Okular

[b***@artifex.com]

http://bugs.ghostscript.com/show_bug.cgi?id=698785

Tor Andersson <***@artifex.com> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |***@artifex.com

--- Comment #2 from Tor Andersson <***@artifex.com> ---
The content stream that contains the graphics drawing operators contains
garbage. Here's a snippet of the beginning of the content stream:

3 0 obj
<<
/Length 1926
/Filter /FlateDecode
stream
1 0 0 1 70.515 594.048 cm
pagesize width 29.7truecm height 21.0truecm
1 0 0 1 -70.515 -594.048 cm
BT
<snip>

As you can see, the 'pagesize', 'width', 'height' and 'true' keywords are
interspliced with some other random text that makes no sense. This looks like
data corruption.

We don't recover in the same way as other interpreters. We keep the 21.0 number
on the stack, which then makes the subsequent 'cm' command have a strange
matrix which messes up the rest of the page.