summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Various "stylistic" improvements suggested by Joshua Bloch's book.David A. Madore2010-04-2214-43/+79
| | | | | | Use @Override annotation. Limit accessibility of fields. Make final what can be. Use complex enums. Don't ignore exceptions. Some more changes on exceptions thrown.
* Navbar.David A. Madore2010-04-192-1/+61
|
* Stray semicolon.David A. Madore2010-04-191-1/+1
|
* Rather than an ad hoc "Options" class, use caller todo directly.David A. Madore2010-04-199-60/+53
|
* Use RuntimeException classes rather than Error to indicate errors.David A. Madore2010-04-194-9/+9
|
* Force encoding on unparsing.David A. Madore2010-04-181-1/+2
|
* Fix misuse of getAttributeNS(), and handling of xml:lang.David A. Madore2010-04-182-15/+23
| | | | | | | | Note to self: getAttributeNS() returns "", not null, when the attribute is nonexistent. Also, the XML spec (now) clearly states that "" is used for an unknown/undefined language.
* Handle <body> element.David A. Madore2010-04-182-0/+78
|
* A handful of meta tags, script and style sections.David A. Madore2010-04-183-0/+72
|
* Rename TodoContext class to Context. Usually call "ctx" its instances.David A. Madore2010-04-188-52/+52
|
* Preliminary handler for <daml> element.David A. Madore2010-04-185-2/+85
|
* Helper class for xml:lang handling.David A. Madore2010-04-182-0/+50
|
* Hide the use of NodeList and NamedNodeMap by real Java ArrayList objects.David A. Madore2010-04-181-9/+25
| | | | | This makes it possible to modify the underlying tree without getting hit by .item(i) returning the wrong next item.
* Use ad hoc classes rather than hashes for context and options.David A. Madore2010-04-177-46/+64
|
* Use ArrayList (supposedly faster because unsynchonized) rather than Vector.David A. Madore2010-04-171-3/+3
|
* Use factories implemented as nested classes (rather than reflected ↵David A. Madore2010-04-174-41/+47
| | | | | | | | | | constructors) to store classes to instantiate. Now every subclass of TodoElement or TodoAttr should also come with a nested subclass of TodoElement.Factory or TodoAttr.Factory which just contains a(n instance) method to call its constructor. In essecne, the factory instance is just used as a pointer to the class to create (or its constructor, as in the previous version).
* Use static initializers correctly (I had forgotten the "static" keyword!).David A. Madore2010-04-162-8/+2
|
* Change dispatching approach: handlers are now part of todoItems.David A. Madore2010-04-1511-106/+130
| | | | | | | Instead of dispatching the todo item in function of the DAML node's local name at todo-handling time, the appropriate todo handler subclass is now instantiated in the todo deque by dispatching the creation of the todo item to the appropriate constructor.
* Instantiate TodoDeque rather than using static variables.David A. Madore2010-04-154-24/+30
| | | | Unfortunately, processing is now rather significantly slower.
* Make handlers instances instead of using static methods.David A. Madore2010-04-146-85/+67
|
* Handle d:xempty attribute, as proof of concept.David A. Madore2010-04-133-7/+22
|
* Framework for handling attributes.David A. Madore2010-04-136-59/+141
|
* Create a todo deque class, and use it.David A. Madore2010-04-133-6/+72
|
* Start writing a todo item handling framework.David A. Madore2010-04-133-3/+53
|
* Start writing code for a todo stack.David A. Madore2010-04-122-0/+26
|
* Purge unnecessary import.David A. Madore2010-04-121-1/+1
|
* Write a semi-decent unparser(=serializer).David A. Madore2010-04-122-13/+74
| | | | Still a bit rough at the edges, but this basically does what I want.
* Move ad hoc serializer to an "Unparser" class (ugly and unfinished).David A. Madore2010-04-122-47/+92
|
* Handle x-daml-magic URL (this should go away eventually, of course).David A. Madore2010-04-111-0/+38
|
* Implement a toy serializer of my own (just the structure).David A. Madore2010-04-111-15/+48
| | | | This should go into a separate class, of course.
* Another attempt at output formatting, this time with serializers. Still sucks.David A. Madore2010-04-111-13/+8
|
* Alternate version using javax interfaces rather than the W3C DOM Load&Save.David A. Madore2010-04-111-19/+25
| | | | | | I was rather hoping this would pretty-print XHTML as required (e.g., put a space before the final slash on empty tags), but alas, such is not the case.
* Start a Java version of a daml engine. So far, just the identity transform.David A. Madore2010-04-112-0/+204
Even for something so completely trivial, this was a pain to write, because some stupid XML catalog was distributed with a DOCTYPE referencing a broken URL which could not be replaced by a local file because the catalog engine was bootstrapping; nor did Xerces provide a way, in org.apache.xerces.util.XMLCatalogResolver, to provide a bootstrap resolver: so I had to partially rewrite XMLCatalogResolver.