I'm trying to embed a subset of a font with the identity encoding range varying
dynamically (only the range used).
For my question, however, let's focus on this specific range which is `a-z`:
fontEncoding = new PoDoFo::PdfIdentityEncoding(0x0061, 0x007A, true);
Assuming you're familiar with creating a font and writing with a painter, I'll
just pseudo code it:
document.CreateFontSubset("Arial", false, false, fontEncoding);
Given this hard-coded range, drawing the text "za" outputs:
<</W
[ 68
[ 443.848
500.0
443.848
500.0
443.848
333.008 ]
74
75
500.0
76
77
277.832
78
[ 500.0
277.832
777.832 ]
81
84
500.0
85
[ 333.008
389.16
277.832 ]
88
89
500.0
90
[ 722.168 ]
91
92
500.0
]
When you view the PDF, the "z" character has an incorrect character spacing,
with a rather large gap after it. If I use the same input string but change the
encoding range max value +1 the same is outputted, except now this is at the
end of the array:
93
[ 443.848 ]
Now the "z" appears correctly spaced. I can open the first broken PDF manually,
add this exact text to the end of the array, save it, and it fixes the "z"
inside it as well. While I'm no PDF spec pro by any means, my understanding is
that this array maps the character indexs to the font glyphs and it appears
that the last entry in this particular range is missing for whatever reason?
(this is likely a wrong assumption, no doubt)
Now, this exactly was with a hardcoded a-z range, but the problem appears with
several ranges I tested arbitrarily and the amount the range is "off" does not
appear to follow any pattern I can decern. For example, if the range ends with
a tilde "~" I need to add 35 to the max range 0x007E.
I can't seem to figure out what I'm doing wrong. If I make the range one less
unit, the last character does not appear (as expected), likewise with the min.
Anyone see some super obvious that I'm overlooking (or abusing)?
Btw, the reason for this need is that I was previously not embedding the font
but was still using the max range of 0xFFFF (per the mailing list suggestion
for unicode support) but the issue we ran into was that as far as I can tell,
this is against spec because doing so creates a CIDToGIDMap which can only be
used if the font is embedded (glyph index values cannot be guaranteed for
TrueType) Section 5.6. For 99% of PDF readers we found this was a null-issue
because they ignore this fact and blindly trust the indexes and usually they
are the same so it works. Google Docs is one implementation that does NOT allow
this and simply doesn't show the text at all. Adobe Reader on iOS-only does
this too.
------------------------------------------------------------------------------
Free Next-Gen Firewall Hardware Offer
Buy your Sophos next-gen firewall before the end March 2013
and get the hardware for free! Learn more.
http://p.sf.net/sfu/sophos-d2d-feb
_______________________________________________
Podofo-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/podofo-users
