[
https://issues.apache.org/jira/browse/PDFBOX-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16815462#comment-16815462
]
Alexey Zhukov edited comment on PDFBOX-4017 at 4/11/19 2:32 PM:
----------------------------------------------------------------
Agreed, 2.0.14 still renders PDF with [^symbol.ttf] from Windows10 glyphs
incorrectly. For example second page of [^test.pdf] contains "Omega" character
(see near "Forward Slope Resistance") from Symbol.ttf and PDFBOX tries to
sequentially search font for codes "Omega", "uni2126" and 8486 with no luck.
While cmap \{3, 0} isn't unicode but symbol it contains high values like 0xF057
etc. Looks like for this case it is better to use font's encoding.inverted map
that maps "Omega" to glyph 87
was (Author: zhukovan):
Agreed, 2.0.14 still renders PDF with [^symbol.ttf] from Windows10 glyphs
incorrectly. For example second page of [^test.pdf] contains "Omega" character
(see near "Forward Slope Resistance") from Symbol.ttf and PDFBOX tries to
sequentially search font for codes "Omega", "uni2126" and 8486 with no luck.
But cmap \{3, 0} isn't unicode but symbol and contains values like 0xF057 etc.
Looks like for this case it is better to use font's encoding.inverted map that
maps "Omega" to glyph 87
> Symbol font glyphs not found on Windows 10 fall creators update
> ---------------------------------------------------------------
>
> Key: PDFBOX-4017
> URL: https://issues.apache.org/jira/browse/PDFBOX-4017
> Project: PDFBox
> Issue Type: Bug
> Components: FontBox, Rendering
> Affects Versions: 2.0.8
> Environment: Windows 10 fall creators update
> Reporter: Tilman Hausherr
> Priority: Major
> Attachments: symbol.ttf, test.pdf
>
>
> Since the Windows 10 fall creators update all files with the Symbol standard
> 14 font are no longer rendered properly. Seems to have something to do with a
> difference in the cmap subtable (all numbers start with F000) and a missing
> postscript table.
> The font has two cmap subtables. The first one has codes that start with 0,
> the second one has codes that start with 0xF000. The code names we get from
> {{getUniNameOfCodePoint(unicodes.codePointAt(0))}} are NOT the low ones, only
> some match, those that are also in ascii.
> For ">", there are two cmap subtables codes: for "Macintosh, Roman" it is
> 0x3e, for "Microsoft, Symbol" it is 0xf03e. I tried forcing the first table
> {{code}}
> cmap = cmapTable.getSubtable(CmapTable.PLATFORM_MACINTOSH, 0);
> {{code}}
> but many codes are missing, e.g. the greek alphabet.
> For "eta" (looks like an "n")
> pdf code 0150 = 0x68
> unicode 0x3b7
> windows code in charmap.exe: 68
> code in subtables: 0x68 and 0xf068
> lozenge
> unicode 0x25ca
> pdf code 0340 = 0xe0
> code in subtables: 0x0e and 0xf0e0
> So it seems one has to bypass this unicode thing and work directly with the
> symbol encoding...
> Websites:
> https://www.microsoft.com/typography/otspec/cmap.htm#windowsPlatform
> https://developer.apple.com/fonts/TrueType-Reference-Manual/RM06/Chap6cmap.html
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
