Finn Bock [Thu, 22 Jan 2004 09:06:05 +0000 (09:06 +0000)]
Abandon codegenerated property makers and use bean-like makers instead.
- ShorthandParsers are made stateless.
- Makers for compound properties inherts from new CompoundPropertyMaker
class and overrides makeNewProperty.
- FOPropertyMapping is no longer generated but is handcoded.
- get() and findProperty() methods are moved from PropertyList to the
property makers.
Finn Bock [Tue, 20 Jan 2004 11:45:48 +0000 (11:45 +0000)]
Fix a NPE caused by avalonization of logging.
Support for 'results' attribute on 'test' element, used by NIST suite.
Explicit close the output file, otherwise the file has a length of 0
when the files are compared (on windows).
Added -d option to enable debug output.
Glen Mazza [Tue, 20 Jan 2004 01:14:33 +0000 (01:14 +0000)]
Interfaces now in alphabetical order and detached from generic interfaces
in autogenerated fo.properties.*; patch from Finn Bock--unneeded method in
PropertyInfo.
Glen Mazza [Mon, 19 Jan 2004 23:56:50 +0000 (23:56 +0000)]
Updated team page--moved Finn, Andreas, Chris and Peter to active status;
set Victor and Keiron to inactive status, removed "former contributor"
section in favor of going back to giving credit within source files.
Glen Mazza [Sat, 17 Jan 2004 19:29:46 +0000 (19:29 +0000)]
Moved the interfaces into the Constants class on a trial basis (we may still
choose to remove them in favor of strictly using Constants.) Two interfaces
(span and position) were removed because of conflicts with other class names.
The interface generation was moved into Constants.xsl, which is run manually
via the XsltToJava ant task and its output is then checked in. As a result,
no more autogeneration of these interfaces at build time will be done.
Glen Mazza [Tue, 13 Jan 2004 23:28:31 +0000 (23:28 +0000)]
static boolean array inheritableProperty[] added, to reduce processing costs
of lookups to see if a property is inheritable. Work based on Finn Bock's patch.
Glen Mazza [Wed, 7 Jan 2004 21:10:38 +0000 (21:10 +0000)]
Last of string->int conversions (more structural improvements still can be
done in PropertyList from Finn's patch, however). Made Constants values
inherent to CompoundDatatype class.
Glen Mazza [Sun, 4 Jan 2004 22:47:36 +0000 (22:47 +0000)]
More String->Int conversion; also removed member variable passing (namespace,
element) within private functions. (Also made more methods private until external
use determined.)
Glen Mazza [Fri, 2 Jan 2004 23:53:09 +0000 (23:53 +0000)]
FOPropertyMapping.GetPropertyId() modified to also be able to return
(base + compound) ID value for a "base.compound" string; more String->Int
conversions in Leader.java.
Glen Mazza [Fri, 2 Jan 2004 22:37:09 +0000 (22:37 +0000)]
Bug 25803 (patch by Simon Pepping): Fix bug occurring when a compound property
is specified (e.g. "leader-length.maximum") before its base property (e.g. "leader-length").
Glen Mazza [Sat, 27 Dec 2003 22:00:38 +0000 (22:00 +0000)]
*Partial* conversion of PropertyList.get(String propName) to new PropertyList.get(int propId) method. This method will remain
overloaded until all calls converted to int constants.
Glen Mazza [Fri, 26 Dec 2003 23:41:47 +0000 (23:41 +0000)]
Property.getPropertyName() switched from returning strings to
integer constants (perh. should be renamed to getPropertyId()?);
change propagated to classes calling this function.
Glen Mazza [Fri, 26 Dec 2003 22:11:17 +0000 (22:11 +0000)]
----------------------------------------------------------------------
Bug #25646 (Patch by Finn Bock): setting SAX Locator (line and column
index of input fo stream) for debugging and better error feedback.
Glen Mazza [Wed, 24 Dec 2003 00:06:14 +0000 (00:06 +0000)]
Removed the elementStringTable HashMap references (never used; was
meant to define element-specific makers for a particular property,
instead of the default make for the property) from fo.FObj and
fo.PropertyList.
Glen Mazza [Mon, 22 Dec 2003 23:23:05 +0000 (23:23 +0000)]
Initial Check-in of PropertySets.java and .xsl. (Similar to Alt-Design's
PropertySets, however uses integer arrays to identify those properties
relevant for an FO.)
Glen Mazza [Mon, 22 Dec 2003 21:50:58 +0000 (21:50 +0000)]
Reversed order of code: add string constant first to HashMap, *then* create
corresponding maker (latter classes weren't having access to the string constant
in their constructors.)
Glen Mazza [Mon, 22 Dec 2003 03:53:31 +0000 (03:53 +0000)]
----------------------------------------------------------------------
More of conversion of strings to integers: Makers now being created via int constants in fo.properties.Constants class.
Temporary string-int conversions currently in Property.java and PropertyList.java, which converts strings from callers into ints for subsequent use.
Glen Mazza [Sat, 20 Dec 2003 17:40:01 +0000 (17:40 +0000)]
Property Makers now being activated via int constants. (Maker classes themselves, as well as code referencing the properties, still need conversion to int's. Also, HashMaps for String and Integer in FOPropertyMapping and FObj temporarily being retained for troubleshooting purposes.) Contribution mainly from Finn Bock (Bug #25480).
Glen Mazza [Mon, 15 Dec 2003 01:07:50 +0000 (01:07 +0000)]
1. Removed the properties.Constants class from automatic generation, and
added Finn Bock's work on Property and Element Constants (currently unused
within application.) The CVS version is now the official one. (The
new constants are not currently being used within the application;
more work is needed.)
2. Build.xml has a new helper XSLTtoJAVA task separate from any of the build
processes to generate the Constants class above (although its output is
expected to fall behind what will be maintained in the CVS version.)
3. Some xml and xsl file renaming to make clearer the purposes of those files.
4. Removed a debug statement from PropertyList.java (forgot to remove from
yesterday's work).
Glen Mazza [Fri, 12 Dec 2003 22:37:39 +0000 (22:37 +0000)]
Moved FontSetup and two helper classes from render.pdf to fonts package.
FontSetup has the PDF fonts as the default, but this class can be extended
if/when another renderer needs its own font setups. (Cannot do this, however,
for AWT's FontSetup at the moment, because its setup() has a different
signature.)
Glen Mazza [Mon, 24 Nov 2003 19:07:35 +0000 (19:07 +0000)]
1.) Chris Bowditch's patch on padding-start implemented. (Changes were
extended to implement part of padding-end as well.)
2.) Switched from TOP/BOTTOM/LEFT/RIGHT to BEFORE/AFTER/START/END
in CommonBorderAndPadding--this is more appropriate for BlockLayout,
also the various places calling it already were either (1) commenting
on the need to switch to the above, or (2) placing the results of the
above functions using the former constants into variables emphasizing
the latter.
Jeremias Maerki [Sat, 22 Nov 2003 19:48:21 +0000 (19:48 +0000)]
New feature for the Ant task: relativebase="true" uses the directory of each FO-file in a fileset as basedir. This is in contrast to the general basedir parameter.
Submitted by: Joe DeVivo <fop.at.joedevivo.com>
1. add method to LayoutStrategy indicating whether an FO Tree should be built for this strategy
2. add logic in Driver.render() to handle the case of a LayoutStrategy that does not want to build an FO Tree