Removed TODO: New validation event to generate an error in case an fo:marker appears as a child of an fo:block-container with absolutely positioned areas.
Additionally: some javadoc updates in BlockContainer.java
Simon Pepping [Thu, 24 Apr 2008 18:19:06 +0000 (18:19 +0000)]
Make the eventResourceGenerator task for the test files a separate
task. Now one can do: ant codegen, compile outside of ant, ant
compile-copy-resources, ant junit-compile-copy-resources.
Partially reverted my revision 641827: The yellow tint of certain JPEG images came from badly extracted ICC color profiles. This is fixed in XML Graphics Commons Trunk (revision 650048).
Simon Pepping [Fri, 18 Apr 2008 19:07:57 +0000 (19:07 +0000)]
An implementation of a positive integer property maker. It implements
convertProperty, which is the correct method to implement if one wants
to add to the make process. This fixes bug 44619.
Introduced PageProductionException and fixed exception throwing for the event related to this exception. It's a RuntimeException descendant as making it a checked exception would have made changes necessary to at least a dozen more classes. Our overall exception style is still not so good.
Split PrintRenderer into PageableRenderer and PrintRenderer (extends PageableRenderer). This allows the use of FOP with JPS. A new example illustrates how this is done.
Removed the code that sets start/end/copies from system properties (now done through renderer options).
Throw a useful exception when the user agent isn't set on the renderer. This has been a problem for more than one person setting up their own renderer using FOUserAgent.setRendererOverride().
Now tunneling all events through a special event listener in tests. This serves two purposes:
- Better readability of the test output.
- Identification of messages that are not sent through the event subsystem, yet.
Bugzilla #41687:
Restored ability to specify from/to and odd/even pages as well as the number of copies for printing from the command-line. Note that this is no longer done via system properties but through the renderer options. Syntax is slightly different. See "fop -print help".
Bugzilla #44678:
Added ability to pass in the PrinterJob instance through the renderer options.
Submitted by: Antti Karanta <Antti.Karanta.at.napa.fi>
Modifications/Additions to patch:
- Don't remove the contructor with the PrinterJob parameter, just deprecate it.
- Adjust old-style print example to use renderer options
First part of the implementation of stage 1 for advanced keeps (see Wiki): Integer values are treated differently from "always" values in keep-together.within-column for all block-level FOs.
Merged branch
https://svn.apache.org/repos/asf/xmlgraphics/fop/branches/Temp_ProcessingFeedback
into Trunk.
Changes on branch:
........
r615153 | jeremias | 2008-01-25 10:07:21 +0100 (Fr, 25 Jan 2008) | 1 line
Created temporary branch for processing feedback.
........
r615155 | jeremias | 2008-01-25 10:11:59 +0100 (Fr, 25 Jan 2008) | 1 line
Initial commit of what I've built already for those who prefer code to minimalistic design docs.
........
r615278 | jeremias | 2008-01-25 18:25:00 +0100 (Fr, 25 Jan 2008) | 1 line
EventProducer interfaces now operational.
........
r615773 | jeremias | 2008-01-28 10:06:16 +0100 (Mo, 28 Jan 2008) | 1 line
No casting in client code when creating EventProducer instances.
........
r616242 | vhennebert | 2008-01-29 11:34:45 +0100 (Di, 29 Jan 2008) | 3 lines
Trick to avoid hard-coding the class name of EventProducer in the source file.
Feel free to revert if it's not ok.
........
r616900 | jeremias | 2008-01-30 21:59:31 +0100 (Mi, 30 Jan 2008) | 1 line
Generate event model XMLs in to the build directory: build/gensrc and build/test-gensrc (the latter is new and needs to be setup as source folder in your IDE!)
........
r616907 | jeremias | 2008-01-30 22:12:59 +0100 (Mi, 30 Jan 2008) | 1 line
Added an XMLResourceBundle that uses an XML file instead of a properties file to load the translations. The XML format is the same as for Cocoon's XMLResourceBundle.
........
r617097 | vhennebert | 2008-01-31 11:53:21 +0100 (Do, 31 Jan 2008) | 2 lines
Minor typo + slight improvement of Javadoc
........
r617176 | jeremias | 2008-01-31 19:14:19 +0100 (Do, 31 Jan 2008) | 5 lines
Renamed FopEvent to Event as suggested by Simon.
EventProducerCollectorTask.java now reads the EventSeverity from a doclet tag.
Added generation of EventProducer translations (including simple merging, no validation, yet)
EventFormatter introduced (only basic functionality, yet).
Added a simple EventListener implementation that uses EventFormatter to convert the events to human-readable, localized messages that are sent to the log via Commons Logging.
........
r617362 | jeremias | 2008-02-01 08:18:07 +0100 (Fr, 01 Feb 2008) | 1 line
Some remaining rename operations based on an earlier discussion.
........
r617413 | jeremias | 2008-02-01 10:46:26 +0100 (Fr, 01 Feb 2008) | 2 lines
Extracted formatting functionality into utility class AdvancedMessageFormat.java.
AdvancedMessageFormat.java now supports conditional sub-groups (delimited by []).
........
r618682 | jeremias | 2008-02-05 17:07:08 +0100 (Di, 05 Feb 2008) | 1 line
Add support for special object formatters (where toString() isn't good enough). ATM, it's hard-coded but could later be hooked into dynamic discovery if we have multiple such formatters. The SAX Locator is the only example for now.
........
r618686 | jeremias | 2008-02-05 17:12:56 +0100 (Di, 05 Feb 2008) | 3 lines
Hooked most of FONode into the new event mechanism. The FOUserAgent provides a DefaultEventBroadcaster instance.
If a producer method declares throwing an exception, the event is automatically marked FATAL and the dynamic proxy throws an exception right after notifying the listeners.
The exceptions are created through the EventExceptionManager. It currently contains only one exception factory for ValidationException. If we need more such factories it's better to register them dynamically. Right now, they're hard-coded.
........
r619313 | jeremias | 2008-02-07 10:14:15 +0100 (Do, 07 Feb 2008) | 1 line
Make sure no events are now just silently swallowed because after upgrading a user doesn't know about the event system.
........
r619314 | jeremias | 2008-02-07 10:14:46 +0100 (Do, 07 Feb 2008) | 1 line
Log what translation file is being written.
........
r619320 | jeremias | 2008-02-07 10:31:00 +0100 (Do, 07 Feb 2008) | 2 lines
FObj hooked into the event system.
Code reduction using a protected method on FONode to acquire a FOValidationEventProducer.
........
r619359 | jeremias | 2008-02-07 11:59:19 +0100 (Do, 07 Feb 2008) | 2 lines
Fop's QName now extends XGCommons' QName to initiate a transition.
Hooked PropertyList into the event mechanism.
........
r631252 | jeremias | 2008-02-26 16:24:33 +0100 (Di, 26 Feb 2008) | 1 line
Removed superfluous warning.
........
r631268 | jeremias | 2008-02-26 17:08:11 +0100 (Di, 26 Feb 2008) | 1 line
Deprecated two methods which are a problem for localization. Also helps finding additional spots to switch over to the event mechanism.
........
r633852 | jeremias | 2008-03-05 15:20:24 +0100 (Mi, 05 Mrz 2008) | 1 line
Added support for additional field styles:
{<fieldname>,if,<true-text>,<false-text>}
{<fieldname>,equals,<test-string>,<true-text>,<false-text>}
........
r633856 | jeremias | 2008-03-05 15:24:04 +0100 (Mi, 05 Mrz 2008) | 2 lines
Javadocs and TODOs.
EventListeners can change the event severity.
........
r633857 | jeremias | 2008-03-05 15:27:08 +0100 (Mi, 05 Mrz 2008) | 4 lines
Javadocs.
Moved out event listener registration into a CompositeEventListener.
Event broadcaster uses the events effective severity, not the initial value (for the case where listeners override the initial value).
Set up a special EventBroadCaster in the FOUserAgent that filters events through a class (FOValidationEventListenerProxy) that adjusts the event severity for relaxed validation.
........
r633858 | jeremias | 2008-03-05 15:32:07 +0100 (Mi, 05 Mrz 2008) | 2 lines
Instead of always decentrally checking whether strict validation is enabled or not, this is now done in a special event listener. The event producer method caller simply indicates whether it can recover from the error condition and continue.
Started switching to event production in table FOs.
........
r634027 | jeremias | 2008-03-05 21:58:35 +0100 (Mi, 05 Mrz 2008) | 7 lines
Moved AdvancedMessageFormat into its own package.
AdvancedMessageFormat got the following added functionality:
- Alternative conditional regions [ bla {field}] -> [ bla {field1}| even more bla {field2}]
- Functions: functions get access to the parameters and they can produce an object that is then formatted ({#gatherContextInfo})
- "if" and "equals" format moved to top-level classes and added by dynamic registration.
EventFormatter now supports includes in the form {{includeName}} so you can include other entries from the resource bundle for better reuse.
Some more events in table code.
........
r634031 | jeremias | 2008-03-05 22:05:22 +0100 (Mi, 05 Mrz 2008) | 1 line
Improved context gathering.
Moved GatherContextInfoFunction to an inner class of FONode to reduce visibilities.
........
r634209 | jeremias | 2008-03-06 11:28:14 +0100 (Do, 06 Mrz 2008) | 1 line
ExceptionFactory is now dynamically registered.
More table warnings and errors switch to events.
........
r634326 | jeremias | 2008-03-06 17:08:16 +0100 (Do, 06 Mrz 2008) | 1 line
Made the "invalidChild" event fully localizable by adding a "lookup" field for the optional rule to be displayed.
And a few switches to the event system.
........
r634692 | jeremias | 2008-03-07 15:31:43 +0100 (Fr, 07 Mrz 2008) | 1 line
More FO tree stuff switched to events.
........
r634712 | jeremias | 2008-03-07 16:19:21 +0100 (Fr, 07 Mrz 2008) | 1 line
Avoid an NPE that says nothing (ex. could happen if the message template is wrong).
........
r634738 | jeremias | 2008-03-07 17:38:21 +0100 (Fr, 07 Mrz 2008) | 2 lines
Non-FO children were not properly run through validation by FOTreeBuilder.
Unified the way that non-FO elements are validated. Some FOs were already fixed. I now fixed the rest, so foreign elements can occur everywhere.
........
r637833 | jeremias | 2008-03-17 12:01:41 +0100 (Mo, 17 Mrz 2008) | 1 line
Exception while cloning for RetrieveMarker to be handled by user as suggested by Andreas.
........
r637835 | jeremias | 2008-03-17 12:03:31 +0100 (Mo, 17 Mrz 2008) | 1 line
Throw a RuntimeException of no other Exception class is specified for an event as a fallback if someone just sets the event severity to FATAL.
........
r637838 | jeremias | 2008-03-17 12:06:10 +0100 (Mo, 17 Mrz 2008) | 1 line
Throw a meaningful exception when the property name is wrong. Otherwise, there will be an ArrayIndexOutOfBoundsException.
........
r637859 | jeremias | 2008-03-17 13:35:26 +0100 (Mo, 17 Mrz 2008) | 1 line
Throw a meaningful exception when the property name is wrong. Otherwise, there will be an ArrayIndexOutOfBoundsException.
........
r637938 | jeremias | 2008-03-17 16:19:51 +0100 (Mo, 17 Mrz 2008) | 1 line
Switched pagination package to events.
........
r637947 | jeremias | 2008-03-17 16:45:16 +0100 (Mo, 17 Mrz 2008) | 1 line
Added support for java.util.text's ChoiceFormat to AdvancedMessageFormat.
Reuse the regexes as constants.
........
r638302 | jeremias | 2008-03-18 11:17:06 +0100 (Di, 18 Mrz 2008) | 1 line
Events for inline-level layout managers.
........
r638774 | jeremias | 2008-03-19 11:17:36 +0100 (Mi, 19 Mrz 2008) | 1 line
Generalized FOValidationEventListenerProxy into FOPEventListenerProxy, the main proxy for FOP's own event manipulation proxy. Done because of support for overflow="hidden" vs. overflow="error-if-overflow".
Switched block-level layout managers to events.
Some cleanup along the way.
........
r639222 | jeremias | 2008-03-20 10:27:34 +0100 (Do, 20 Mrz 2008) | 2 lines
Some initial work for event forwarding from Batik.
Missing errors/exceptions converted to events in PageSequenceMaster.
........
r639270 | jeremias | 2008-03-20 13:50:35 +0100 (Do, 20 Mrz 2008) | 1 line
Removed DEBUG event severity again. Promoted constrained geometry adjustment event to INFO level as per discussion.
........
r640395 | jeremias | 2008-03-24 13:39:13 +0100 (Mo, 24 Mrz 2008) | 3 lines
Moved the creation of the fallback LoggingEventListener to FOUserAgent so event before the startDocument() SAX event arrive in the log.
Dynamic discovery of event models. Renderers and extensions can register renderer-specific event models.
Switched the most important parts of the renderers to events (maybe not everything is converted).
........
r640397 | jeremias | 2008-03-24 13:43:04 +0100 (Mo, 24 Mrz 2008) | 1 line
Remaining fixcrlfs. Xalan likes to mix CRLF and LF on Windows.
........
r640398 | jeremias | 2008-03-24 13:43:54 +0100 (Mo, 24 Mrz 2008) | 1 line
Event in area package.
Exposed getUserAgent() in Renderer interface (was already public in AbstractRenderer).
........
r642972 | jeremias | 2008-03-31 14:18:39 +0200 (Mo, 31 Mrz 2008) | 1 line
Plugged fonts package into the event subsystem. Note: I did not follow the same pattern as for the rest as the font package is to be considered FOP-external, so I just added a manual adapter for the FontEventListener. This demonstrates how an external library can be integrated with the event system.
Missing warning for unknown formatting objects added. Warning is not issued by the ElementMappingRegistry anymore but by FOTreeBuilder which has access to more context information.
........
r642997 | jeremias | 2008-03-31 16:10:08 +0200 (Mo, 31 Mrz 2008) | 1 line
Added an example to demonstrate how to write your own event listener and how to deal with the exceptions thrown in the process.
........
r642998 | jeremias | 2008-03-31 16:13:40 +0200 (Mo, 31 Mrz 2008) | 1 line
Removed unused method. Event formatting should not be part of the Event class. Use EventFormatter.format(Event) instead.
........
r643066 | jeremias | 2008-03-31 19:18:54 +0200 (Mo, 31 Mrz 2008) | 1 line
First part of the event subsystem documentation (DRAFT).
........
r643784 | jeremias | 2008-04-02 10:05:33 +0200 (Mi, 02 Apr 2008) | 1 line
More documentation.
........
r643785 | jeremias | 2008-04-02 10:06:38 +0200 (Mi, 02 Apr 2008) | 1 line
Some nits.
........
r643787 | jeremias | 2008-04-02 10:24:41 +0200 (Mi, 02 Apr 2008) | 1 line
Another small typo fix
........
r647678 | jeremias | 2008-04-14 09:20:26 +0200 (Mo, 14 Apr 2008) | 1 line
Renamed *EventProducer.Factory.create() to *EventProducer.Provider.get() to better reflect what the method does (instances may be cached and reused).
........
Initialized merge tracking via "svnmerge" with revisions "1-615152" from
https://svn.apache.org/repos/asf/xmlgraphics/fop/branches/Temp_ProcessingFeedback
Bugzilla #44744:
Disable drawString(AttributedCharacterIterator, float, float) as it contains bugs and rely on the fallback implementation from AbstractGraphics2D.
Bugzilla #44737:
Added support for auto-configuring TrueType Collections. XML font metrics files for *.ttc fonts are not required anymore.
Submitted by: Jason Harrop <jason.at.plutext.org>
Changes to patch or in addition to the patch:
- Tab chars removed and Checkstyle issues fixed
- Some simplifications in the cache handling (CachedFontInfo is obsolete and less cache-private information is exposed to the outside).
- TTCs are fully detected and registered with FOP.
- TTCs can also be registered using a "font" element. The new "sub-font" attribute selected the sub-font in the TTC.
- Bug fixed in TTFFile: Font names were not decoded correctly (ex. font names in Chinese)
- Minimal docs.
Jeremias Maerki [Thu, 27 Mar 2008 14:29:44 +0000 (14:29 +0000)]
When a JPEG image is embedded, an optionally embedded color profile is filtered out as it's already embedded separately in the PDF file.
Worked around a problem (PDF renderer) with JPEG image containing RGB color profiles which are not sRGB. The images drifted into yellow. The color profile is simply disabled in this case. Please let us know if you know what the problem could be.
Jeremias Maerki [Thu, 27 Mar 2008 08:49:41 +0000 (08:49 +0000)]
Added support for addressing all glyphs available in a Type 1 font, not just the ones in the font's primary encoding.
Typeface: getEncoding() changed to getEncodingName() to make clearer what is held here.
Some cleanup in the font classes to put the various things in more appropriate places.
Created a common base class for all Base 14 fonts (makes the hierarchy clearer).
Made PDFTextUtil more universally useful and made use of it in PDFRenderer, too.
Made PDFStream.add(String) more efficient. The encoding converter is not called for each invocation anymore as the whole thing get buffered by a BufferedWriter (as suggested by the javadoc of OutputStreamWriter).
Cleanup and minor refactoring:
* consolidate addId() in AbstractLayoutManager
* replace getPSLM().addIdToPage(getXXX().getId()) in various LMs
Smaller changes/cleanup in the affected LMs include:
* removal of some redundant casts
* simplified conditionals
* avoid duplicating the reference to the FObj as much as possible:
the reference is already stored in AbstractLM, yet every subclass seems to
add another reference...
Updated copyright years
........
r637075 | vhennebert | 2008-03-14 12:19:45 +0000 (Fri, 14 Mar 2008) | 4 lines
Bugzilla #44412:
Bugfix: When there was a forced break after a block with (conditional) borders the border-after wasn't painted.
Changes made by Jeremias
........
r637119 | jeremias | 2008-03-14 14:41:03 +0000 (Fri, 14 Mar 2008) | 2 lines
Bugzilla #44412:
Bugfix: The before border of a block is no longer swallowed if its first child issues a break-before.
........
r637791 | jeremias | 2008-03-17 08:59:51 +0000 (Mon, 17 Mar 2008) | 3 lines
Removed old image adapters.
Removed support for Java 1.3 building as it didn't work anymore anyway.
Added a check to require Java 1.4 for building.
........
r637857 | jeremias | 2008-03-17 12:24:33 +0000 (Mon, 17 Mar 2008) | 3 lines
Partially reverted revision 637791:
Restored JAI as a required dependency for distribution builds (used for error diffusion dithering in the PCL Renderer). It is optional otherwise.
Updated the licensing info and release notes to make the changes clear.
........
r637993 | vhennebert | 2008-03-17 17:48:44 +0000 (Mon, 17 Mar 2008) | 3 lines
Bugzilla #44621: when the after border of a cell, in the trailing case, is bigger than in the normal case, the generated sequence of Knuth elements was wrong, leading to content being swallowed
This is a partial fix only: the content is no longer swallowed, but the penalty is wrong (shorter than it should be). This will lead to the table overflowing the region-body without warning.
........
r638048 | vhennebert | 2008-03-17 20:00:36 +0000 (Mon, 17 Mar 2008) | 2 lines
Changed the documentation to reflect the move to Java 1.4 as a minimum requirement
........
r638308 | vhennebert | 2008-03-18 10:26:32 +0000 (Tue, 18 Mar 2008) | 2 lines
Simplified the getNextKnuthElements method a little bit
........
r638316 | vhennebert | 2008-03-18 11:06:59 +0000 (Tue, 18 Mar 2008) | 2 lines
Removed the 'jdk1.4' suffix appended to the binary artifacts. There's no point in making the distinction anymore since there is only one binary now
........
r638317 | vhennebert | 2008-03-18 11:08:29 +0000 (Tue, 18 Mar 2008) | 2 lines
Reverting changes of rev. 637057. I applied the changes to the 0.95 branch instead so that they make their way into the 0.95 release. I will merge them back to the Trunk via svnmerge
Updated the disabled-testcases.xml file: re-enabled testcases which could easily be updated, removed the other ones which anyway were superseded by more complete ones
Jeremias Maerki [Mon, 10 Mar 2008 10:06:37 +0000 (10:06 +0000)]
Fixed NPE in BlockContainerLayoutManager when used as a child of an inline-level FO. Split IP and BP stack limits in LayoutContext (there's now a certain amount of redundancy with "refIPD" in LayoutContext which I didn't resolve).
Areas are now generated for block-level FOs when used as children of inline-level FOs.
ClassCastException in ListLayoutManager.mustKeepTogether() fixed (occured if used as child of an inline-level FO).
- Fixed the rendering of the area of a table corresponding to border-separation, which must be filled with the background of the table, and not the rows.
- Improved conformance: even if a table-cell spans several rows its background must correspond to the first row spanned.
- Added support for background on fo:table-column and fo:table-header/footer/body
TODO more testcases needed, especially tables with collapsing border model and tables broken over pages
Jeremias Maerki [Tue, 4 Mar 2008 17:35:52 +0000 (17:35 +0000)]
Update FOP poster for OpenExpo 2008 in Bern.
This fixes various warnings, updates the history and simplifies configuration (if you're on Windows, it will work out of the box because all fonts are available).
Jeremias Maerki [Tue, 4 Mar 2008 17:30:40 +0000 (17:30 +0000)]
Bugfix: If there's shrink available in an auto-height situation, the difference is now set to 0 in order to avoid unwanted effects. IOW, never shrink in an auto-height situation.
Jeremias Maerki [Tue, 4 Mar 2008 11:50:34 +0000 (11:50 +0000)]
Renamed to 0.95beta.
Made changes and release notes work again by switching to the projectInfo plug-in rather than relying on our local copy of the stylesheets. There's a little bug (missing images) in the release notes for which I submitted a patch (https://issues.apache.org/jira/browse/FOR-1076) to Forrest.
Brought status.xml more in line with the official status DTD.
Jeremias Maerki [Tue, 4 Mar 2008 09:15:50 +0000 (09:15 +0000)]
Using Forrest Trunk from now on (in order to use FOP 0.94 for PDFs). We'll switch to a stable release as soon as a release with a new FOP version is available.
Jeremias Maerki [Mon, 3 Mar 2008 07:51:14 +0000 (07:51 +0000)]
Tried to make font configuration easier to understand by splitting the basics from the advanced stuff. Otherwise, people will believe for the next 100 years that generating XML font metrics file is still always necessary.
Jeremias Maerki [Fri, 29 Feb 2008 14:03:01 +0000 (14:03 +0000)]
Bugfix: Disabled setting widths from AFM since we're using WinAnsiEncoding internally but the AFM delivers the character widths using the font's native encoding which causes problems with some characters.
Jeremias Maerki [Tue, 26 Feb 2008 11:07:20 +0000 (11:07 +0000)]
Reverted revision 603590 which broke the AFP renderer. The other changes after revision 603590 should be preserved as well as possible, including the switch to the new image library.
Adjustments for change in behaviour of the block viewport CTM.
Bugfix: SVG painting is now always done on a color canvas (like for PCL) because Batik cannot handle gradients on a grayscale canvas.
- added full support for keep-with-previous on table-row and in table-cell
- added more testcases for keeps in tables
- assign the right table-row element to grid units that are not on the first row spanned
- slightly re-worked the RowGroupBuilder interface to make it more SAX-like
Jeremias Maerki [Tue, 19 Feb 2008 15:47:48 +0000 (15:47 +0000)]
Added support for reading the OS/2 table's usWeightClass value which supports the same font weight values as we use in XSL-FO.
However, in my tests these values proved to be unreliable (like ExtraBlack fonts returning 400). I just hooked the whole thing in so this work isn't lost if anyone has an idea to make it work. The FontInfoFinder will continue to only use guessed font weights for now.
Jeremias Maerki [Tue, 19 Feb 2008 13:08:39 +0000 (13:08 +0000)]
Bugzilla #44451:
The patch fixes two problems:
- If distinct fonts declare the same font family name, the resulting font mapping is currently arbitrary; now, a name-similarity heuristic is used to prioritize the font mappings.
- "Medium" and "demi" fonts are now recognized as "bold", solving several real-world problems (although this solution may be an oversimplification).
Submitted by: Justus Piater <justus-bulk.at.piater.name>
Patch modified by jeremias:
- Style fixes (tab chars)
- Refined font weight identification: medium -> 500, semi/demi -> 600
projects / xmlgraphics-fop.git / log