[
https://issues.apache.org/jira/browse/PDFBOX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12836416#action_12836416
]
Villu Ruusmann commented on PDFBOX-624:
---------------------------------------
This is a Type1C font issue. Namely, TeX has generated font programs which
encode the "width" operand of "hsbw" and "sbw" operations as a 5-byte number
(the value is typically in range from 300 to 1000, and other software packages
encode it as a 2- or 3-byte number). To make things even more complicated, TeX
has stored the value in the highest 2 bytes (ie. bitmask 0xffff0000), not in
the lowest 2 bytes (ie. bitmask 0x0000ffff).
I have prepared a small patch that makes TeX generated PDF documents usable.
However, this is a hack and may cause problems elsewhere (my own text
extraction tests pass successfully).
> Misplaced text
> --------------
>
> Key: PDFBOX-624
> URL: https://issues.apache.org/jira/browse/PDFBOX-624
> Project: PDFBox
> Issue Type: Bug
> Components: FontBox, Text extraction, Utilities
> Affects Versions: 1.0.1
> Reporter: Villu Ruusmann
> Priority: Critical
> Attachments: documenta_math.pdf, documenta_math.txt,
> documenta_math_page4.png
>
>
> Thomas Fischer reported to [email protected] that
> org.apache.pdfbox.ExtractText interchanges typographic ligatures "fi" and
> "fl". The sample document "documenta_math.pdf" was created using TeX and AFPL
> Ghostscript 6.50.
> I used PDFBox 1.0.1-SNAPSHOT to verify this problem. The "fi" ligature
> behaves correctly (ie. text extraction yields "finite" and "infinite", not
> "flnite" and "inflnite"), but the overall text layout is a complete mess.
> Please see the PDF text extraction result "documenta_math.txt" and PDF
> rendering result "documenta_math_page4.png".
> The cause of the horizontal text misplacement is not yet known. This could
> affect all PDF documents which have been created using TeX.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
