[
https://issues.apache.org/jira/browse/PDFBOX-2533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14230075#comment-14230075
]
Tilman Hausherr commented on PDFBOX-2533:
-----------------------------------------
The cause is a problem with decryption. The decryption in the non-sequential
parser doesn't have recursion, while the old one uses the securityHandler,
which does have recursion. Two attributes in CIDSystemInfo (Ordering and
Registry) are not decrypted.
{code}
712 0 obj<</Subtype/CIDFontType2/FontDescriptor 711 0
R/BaseFont/Batang/W[1[333]4[623]9 10 376 13[291 625 291 375 595]18 26 595 27 28
333 30[625]34[736 708]36[708 734 671 644 740 765 326 425]45[638 894 775 736
654]51[671 625 750 775 750 945 687 685 625]66[541 576 541 576]70[576 365 583
574 289]76[572 291 845 570 583 574]82[574 439 519 359 566 572 808 605 587
500]130[812]499[485]511[522 541]]/CIDSystemInfo<</Supplement 1/Ordering([
}÷Çè)/Registry(Q+`ðÃ)>>/DW 1000/Type/Font>>
endobj
{code}
So either we'd have to add another level, or implement a decryptDictionary
method, or use the one from SecurityHandler. (There's probably a reason that we
don't...)
> Poor rendering with non-sequential parser
> -----------------------------------------
>
> Key: PDFBOX-2533
> URL: https://issues.apache.org/jira/browse/PDFBOX-2533
> Project: PDFBox
> Issue Type: Bug
> Components: Parsing
> Affects Versions: 1.8.8, 2.0.0
> Reporter: Tilman Hausherr
> Attachments: 905605.pdf
>
>
> p39 + 40 of the attached file are poorly rendered with the non-Sequential
> parser. With the old parser, while the glyphs are not perfect, they represent
> the correct character. This is for 1.8.8.
> In 2.0 the rendering is incorrect, but I didn't test with the old parser
> because that one is no longer available from the command line.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
