[OpenJDK 2D-Dev] [9] Review request for 8023990: regression : postscript file size increase from 6u17

anton nashatyrev anton.nashatyrev at oracle.com
Tue Jan 28 18:52:46 UTC 2014


Hello,

     please review the following fix for JDK-9.

Phil, Andrew,

     the fix had been approved by you for JDK-8 (but not yet pushed), 
this is the 'backport' to JDK-9. The patch has been applied without any 
modifications. The fix for JDK-9 was tested with the bug testcase and 
Font2D demo.

bug: https://jbs.oracle.com/bugs/browse/JDK-8023990
fix for JDK-8: http://cr.openjdk.java.net/%7Ealitvinov/8023990/webrev.00
fix for JDK-9: http://cr.openjdk.java.net/~anashaty/8023990/webrev.00/ 
<http://cr.openjdk.java.net/%7Eanashaty/8023990/webrev.00/>

[description from original review]

     When the FcFontConfiguration was added for using fontconfig on a 
set of Linux platforms (as well as OpenSolaris) (the corresponding rfe 
is here <https://bugs.openjdk.java.net/browse/JDK-8023990>), the 
implementation of the method FcFontConfiguration.getFontDescriptors() 
returned just an empty array. This functionality is actually not used by 
anyone except of the PSPrinterJob which tries to find out whether the 
string could be represented using PS embedded font. If this procedure 
fails then the PSPrinterJob falls back to glyphs outline printing. As I 
understand this normally happened for strings containing non-ASCII 
characters. For ASCII strings PSPrinterJob usually selected the PS 
embedded font. With the addition of FcFontConfiguration even ASCII 
strings output starts to fall back to outlines.

     The fix is some basic implementation of the getFontDescriptors() 
functionality. The returned FontDecriptors don't contain missing glyph 
information. Though it might be queried from the fontconfig library 
(with some additional effort) this doesn't change the behavior of 
PSPrinterJob, since all the fonts has the ISO 10646 encoding (UTF-8), 
and the PSPrinterJob will still be falling back to outlines for any 
non-ASCII strings.

Thanks!
Anton.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/2d-dev/attachments/20140128/ff760a46/attachment.html>


More information about the 2d-dev mailing list