--- Comment #5 from Jonathan Levinson <> 2009-09-21 
10:19:15 PDT ---
Hi Vincent,

It was my pleasure to help out.  Thank you for your review of the problems you
think still remain in the font shorthand code.  

When you closed 47859, did you noticed I had attached a patch that fixes one of
the issues you mention below:

"Also, it doesn't
seem to handle the case where there are more than one comma (i.e., at least 3
families specified)."

This patch, which (I believe) fixes the problem you mention, was attached to
47859, which you closed as a duplicate. 

I infer from reading and studying the code that the purpose of the complicated
""while (!fontFamilyParsed)" loop is to identify the beginning of the font
family token.  The FontFamilyProperty parser assumes it has been given a string
which represents a font family and the font shorthand routine is trying to
identify the font family token (a string) which is sent to FontFamilyProperty. 
The font shorthand code is not trying to do any parsing.  I say this with all
due respect for your greater experience with Apache FOP.  

You've made other comments I'll have to investigate before replying further.  I
need to add more test cases to verify that my fix (the one in 47859) corrects
all issues with font shorthand property.

I can see your point about the desirability of a font shorthand parser. 
Currently the font short hand code tries to identify the strings which should
be sent to each of the font property parsers such as FontFamilyProperty.  So
the font shorthand code is a tokenizer rather than a parser.  Its goal is to
identify the tokens and then send the tokens to property parsers for further
processing.  It could be a tokenizer is all this really needed, but I don't
know at this point.

Best Regards,
Jonathan S. Levinson
Senior Software Developer
Object Group

Configure bugmail:
------- You are receiving this mail because: -------
You are the assignee for the bug.

Reply via email to