Tilman, We totally understand not wanting to expose more of the inner workings than is necessary.
We did look at DrawPrintTextLocations and as far as we can see the core code in both is similar (DrawPrintTextLocations.calculateGlyphBounds vs Glyph.getPathForCharacterCode) but the the Glyph version is more complete / catches more edge cases. We are hoping to open source the work we've done as library on top of PDFBox, so perhaps that's an opportunity to pull aspects back into PDFBox if they fit. Thanks for the response (sorry I wasn't subscribed so can't reply directly, I thought I'd get a direct email). C On Thu, Feb 2, 2017 at 9:08 AM, Chris Flatley <[email protected]> wrote: > Hi, > > We've been working to extract better text from some specific PDFs. Due to > ordering, sizing, etc of characters, we've found the need to use the font > glyph information as part of our calculations. > > In PDFBox these (Glyph2D, CIDType0Glyph2D, TTFGlyph2D, Type1Glyph2D) are > package level. > > Would you have any problems in making these classes and constructors > public available? > > I can raise a JIRA ticket with patch (obviously only 8 or so lines). > > Many thanks, > > C >
