[
https://issues.apache.org/jira/browse/PDFBOX-2372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14147304#comment-14147304
]
John Hewson edited comment on PDFBOX-2372 at 9/25/14 3:08 AM:
--------------------------------------------------------------
This is really tricky. I spent hours experimenting with Acrobat and discovered
that the only way to determine whether GID 0 would be drawn was Acrobat
Preflight's "base 14" font indicator (better known as "Standard 14" fonts).
It's not enough for a font to be non-embedded and use one of the Standard 14
names, it must also use an appropriate Encoding. Some trial and error
experimenting shows that the encoding must be non-symbolic, and must not use
Differences.
For reference, here's what I've determined about the relevant non-embedded
fonts in the test files:
||File||Type||BaseName||Encoding||Symbolic||Acrobat Preflight "Base
14"||Renders .notdef (GID0)||
|PDFBOX-563-acroform.pdf | Type 0 / CIDType2 | "Arial" | Identity-H | Yes | No
| No |
|PDFBOX-1735-confidential.pdf | TrueType | "LucidaConsole" | WinAnsiEncoding |
No | No | No |
|PDFBOX-2245-052567.pdf | TrueType | "TimesNewRoman" | WinAnsiEncoding +
Differences | No | No | No |
|PDFBOX-2334.pdf | TrueType | "Arial" | StandardEncoding | No | Yes|Yes|
was (Author: jahewson):
This is really tricky. I spent hours experimenting with Acrobat and discovered
that the only way to determine whether GID 0 would be drawn was Acrobat
Preflight's "base 14" font indicator (better known as "Standard 14" fonts).
It's not enough for a font to be non-embedded and use one of the Standard 14
names, it must also use an appropriate Encoding. Some trial and error
experimenting shows that the encoding must be non-symbolic, and must not use
Differences.
For reference, here's what I've determined about the relevant non-embedded
fonts in the test files:
||File||Type||BaseName||Encoding||Symbolic||Acrobat Preflight "Base
14"||Renders .notdef (GID0)||
|PDFBOX-563-acroform.pdf | Type 0 | "Arial" | Identity-H | Yes | No | No |
|PDFBOX-1735-confidential.pdf | TrueType | "LucidaConsole" | WinAnsiEncoding |
No | No | No |
|PDFBOX-2245-052567.pdf | TrueType | "TimesNewRoman" | WinAnsiEncoding +
Differences | No | No | No |
|PDFBOX-2334.pdf | TrueType | "Arial" | StandardEncoding | No | Yes|Yes|
> Regressions 19.9.2014
> ---------------------
>
> Key: PDFBOX-2372
> URL: https://issues.apache.org/jira/browse/PDFBOX-2372
> Project: PDFBox
> Issue Type: Bug
> Components: FontBox
> Affects Versions: 2.0.0
> Reporter: Tilman Hausherr
> Assignee: John Hewson
> Labels: regression
> Fix For: 2.0.0
>
>
> There are several regressions from the changes done on the evening of
> 19.9.2014. Because I can't map these to one single change, I had to open a
> new issue.
> PDFBOX-563-acroform.pdf: trash glyphs at the bottom
> PDFBOX-1735-confidential.pdf p7: trash glyphs on the left
> PDFBOX-2245-052567.pdf: trash glyphs
> PDFBOX-2251-070075.pdf: trash glyphs
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
