[
https://issues.apache.org/jira/browse/PDFBOX-1895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13899364#comment-13899364
]
John Hewson commented on PDFBOX-1895:
-------------------------------------
{quote}
NTT Docomo would not issue their 3rd quarter financial report in a broken format
{quote}
The PDF file is broken, that's why Acrobat says "The file is damaged" when
opening it. It is missing required embedded fonts which use an Identity
CIDToGID mapping, meaning that it is not safe to substitute the font. The
CIDSystemInfo it contains specifies the registry "älUx" with ordering "ásVsó"
which should be "Adobe" and "Japan1". There may be other problems too.
The odd thing is that missing fonts with an Identity CIDToGID mapping cannot be
substituted, because a GID is an internal, font-specific detail and can differ
between two otherwise identical fonts. Yet Adobe Reader is somehow able to
perform this substitution. For example, on my machine File > Properties > Fonts
tells me that MS-Mincho has been substituted for HiraMinPro-W6, a Type 1 font!
I can only think that Adobe has some sort of non-standard heuristic which is
being used to safely substitute the font. Whatever changes PDFBox makes when it
removes the signature, one of them is causing the heuristic to fail to identify
that the missing fonts can be safely substituted.
> Font definitions must precede font references
> ---------------------------------------------
>
> Key: PDFBOX-1895
> URL: https://issues.apache.org/jira/browse/PDFBOX-1895
> Project: PDFBox
> Issue Type: Bug
> Components: Writing
> Affects Versions: 1.8.3, 1.8.4
> Reporter: Pat Hickey
>
> When re-writing a document with font descriptions, Adobe Reader is unable to
> display the fonts in the document. Reader can display the fonts in the
> original document. The difference is that in the original document, the font
> descriptions are in lower object numbers than the font references; in the
> output document, the font descriptions are in higher object numbers than the
> font references. Is there a quick way to re-order them?
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
