2001-11-24 Jim Menard * tests/testresource.rb: changed to reflect new external id classes. * tests/tokenizertester.rb: changed to reflect new external id classes. * nqxml/tokenizer.rb (nextDoctype): create external ids. (nextExternalId): factored out of new code. (nextDeclSep): fixed error messages to contain '%' instead of '&'. * nqxml/entities.rb: added new classes ExternalID, SystemExternalID, and PublicExternalID. (Doctype.writeXMLTo): call externalId.to_s when writing. (EntityTag): constructor and attribute changed to hold external id as well as text. (GeneralEntityTag): changed to hold either entity value or external id and NDATA decl. (ParameterEntityTag): changed to hold either entity value or external id. (Element): superclass changed from EntityTag to NamedEntity because of change in behavior and state of EntityTag. Remaining args stored in string argString. (Attlist): same as Element. (Notation): same as Element. 2001-11-22 Jim Menard * tests/oasis.rb: added this file, courtesy of TAKAHASHI Masayoshi. 2001-11-20 Jim Menard * tests/test.rb: run all tests in one big bunch instead of a number of small bunches. 2001-11-19 Jim Menard * Version 1.1.2 released. * tests/testresource.rb (initialize): changed doctype to PUBLIC to test new code. Added eight-bit char to text data. * nqxml/entities.rb (XMLDecl.writeXMLTo): write version, encoding, and standalone attributes (if they are defined) in order because order is significant. * nqxml/tokenizer.rb: removed '\xd' char from NOT_SPACES_REGEX and PUBLIC_ID_LITERAL_REGEX because that char is removed from input stream anyway. One fix and two mods to NAME_REGEX: I forgot to backslash '.' so anything was matching; now accepts all legal 8-bit chars, not just 7-bit; changed anchor from '^' to '\A'. Added ENCODING_NAME_REGEX. (nextDoctype): save the quotes around public id literal. (restOfXMLDecl): use ENCODING_NAME_REGEX to check for legal "encoding" attribute name chars. Make sure "standalone" value is either "yes" or "no". (everywhere): made Input code inline, for speed. Makes code bigger but faster. 2001-08-13 Jim Menard * Version 1.1.1 released. 2001-08-10 Jim Menard * install.rb: use File.join() instead of explicitly using slashes to create install file names. * docs/sgml2text.rb: added "command" to list of tags that need deletion when writing text file. * docs/index.html: fixed links to download pages for other projects. * tests/writertester.rb (compare_test_prettify): take new indentation into account when testing prettify code. Add test for empty tag. 2001-08-06 Jim Menard * nqxml/writer.rb (attribute): call to_s on the attribute value. Indent tags when prettifying. (endElement): Minimize empty tags (for example, write "" instead of ""). * Version 1.1.0 released. 2001-07-24 Jim Menard * release.rb: changed to reflect changes in docs directory (see next item). * docs/*: moved to using LaTeX instead of DocBook. I don't like Red Hat's docbook2* tools and style sheets at all. All of the old files are in the new docs/docbook directory. 2001-07-23 Jim Menard * nqxml/tokenizer_old.rb (initialize): removed unused attribute parseLevel. 2001-07-19 Jim Menard * tests/*.rb: created dispatchtester.rb. Modified code to use list of class names (no good reason, really). Moved some require statements into testing classes. 2001-07-18 Jim Menard * tests/*.rb: changed references to ProcessingInstruction to XMLDecl. Added new ProcessingInstruction test case data. * nqxml/treeparser.rb (miscEntity?): added XMLDecl to list of misc entities. * nqxml/streamingparser.rb (each): modified to know about the new XMLDecl class. * nqxml/entities.rb: ProcessingInstruction made subclass of NamedEntity, not NamedAttributes. New class XMLDecl (subclass of NamedAttributes) created. * nqxml/tokenizer.rb (nextPublicIdLiteral): stopped returning quotation marks around public id literal. Fixed error message at top. (nextProcessingInstruction): modified to call either restOfXMLDecl or restOfProcessingInstruction based on name seen after ' * nqxml/treeparser.rb (initialize): moved up to top of file; why on earth did I ever put it at the bottom? 2001-07-16 Jim Menard * nqxml/document.rb (writeTo): added prettify parameter; write newlines when appropriate. * docs/README.sgml: added prettify explanation to text and examples for writer. * nqxml/writer.rb: added prettify boolean instance variable, added new constructor argument to set it, and use it when writing '>' at end of all tags (both start and end). 2001-07-13 Jim Menard * nqxml/tokenizer.rb (nextDeclSep): fixed test for undefined parameter entity replacement text. 2001-07-12 Jim Menard * nqxml/tokenizer.rb: changed Input instance variables @string, @length, and @replaceRefs to read-only (attr_reader instead of attr_accessor). 2001-07-10 Jim Menard * docs/README.sgml: a new section describes my plan to start deleting older release versions from the Web site. * nqxml/tokenizer.rb (initialize): changed line break substitution to explicitly replace "\r\n" and "\r" with "\n". The old code, which replaced $/ with "\n", was incorrect because the input XML file may use line breaks other than $/. * tests/tokenizertester.rb (test_tokens): removed nop line of code. (compare_tokens_with_expected): added to hold common to test_tokens() and test_different_newlines(). 2001-07-09 Jim Menard * docs/index.html: tweaked navigation bar on left-hand side of page. 2001-06-29 Jim Menard * nqxml/tokenizer.rb (peekMatches?): removed alias :peekMatches because it is not used anywhere in the code. (peekChar): removed alias :peekChars and removed argument because neither is used in the code anywhere. 2001-06-25 Jim Menard * Version 1.0.2 released. * tests/writertester.rb (compare_to_rsrc_xml): strip some whitespace from around the expected DOCTYPE's prolog elements. * nqxml/tokenizer.rb (nextDoctype): put quotes around quoted literals in PUBLIC and SYSTEM external ids. (nextPublicIdLiteral): keep quotes around quoted literal. * tests/testresource.rb: make sure doctype's SYSTEM public id has quotes around the value. 2001-06-19 Jim Menard * Version 1.0.1 released. * nqxml/entities.rb (Doctype): added writeXMLTo method. 2001-06-18 Jim Menard * Version 1.0.0 released. * nqxml/info.rb: finally calling this release 1.0.0 (instead of 0.6.4). Released 6/18/2001. 2001-06-14 Jim Menard * nqxml/entities.rb: explicitly convert entity attribute values to strings on the way out. 2001-06-13 Jim Menard * docs/README.sgml: email address moved to correct place in tag. 2001-06-08 Jim Menard * nqxml/writer.rb: added some comments to the code. * nqxml/utils.rb: added some comments to the code. * nqxml/domparser.rb: deleted from source tree; use NQXML::TreeParser instead. (DOMParser has been an alias to TreeParser for a long time, and the docs warned that this would happen.) 2001-05-30 Jim Menard * docs/index.html: added link to another of my Ruby releases, RuBoids. 2001-05-25 Jim Menard * docs/README.sgml: links to example scripts didn't work on Web site; they do now because I added "../" before each link. Fixed an example script link that pointed to the wrong script. Made a few minor language tweaks. Fixed a few spelling errors. 2001-05-17 Jim Menard * Version 0.6.3 released. * release.rb: there is no reason to ``make clean'' in the docs directory, so we don't do it anymore. * nqxml/entities.rb (Comment.initialize): regex that grabs text from comment now spans newlines. * tests/tokenizertester.rb (test_comment_with_newline_bug): added this routine to test bug reported by Dan Munk. * nqxml/tokenizer.rb (nextDoctype): when comment start found, skip leading dashes. (nextBangTag): ditto. (nextComment): slurped text no longer includes leading dashes. 2001-05-16 Jim Menard * nqxml/document.rb (initialize): set @doctype and @rootNode to nil. * Version 0.6.2 released. * nqxml/tokenizer.rb: changed ``+='' to ``<<'' when concatenating values onto the end of a string. This is more efficient. * nqxml/entities.rb: changed ``+='' to ``<<'' when concatenating values onto the end of a string. This is more efficient. 2001-04-24 Jim Menard * docs/README.sgml: added missing backslashes before newline chars in example code. 2001-04-16 Jim Menard * nqxml/entities.rb (isTagStart): created. Also added aliases :tagStart? and :tagEnd?. * Version 0.6.1 released. 2001-04-10 Jim Menard * docs/sgml2text.rb: substitute '&' after all the other predefined entities so we don't turn '<' into '<'. * nqxml/tokenizer.rb (textUpTo): use regexp instead of array. (nextText): use regexp instead of array in call to textUpTo. Added comment explaining why eof check must come before peekChar call. (normalizeAttributeValue): tweaked for much better performance. 2001-04-09 Jim Menard * nqxml/tokenizer.rb general performance enhancements (cache current input and input's string length). (nextQuotedLiteral): moved check for '<' into normalizeAttributeValue(). (normalizeAttributeValue): check for '<' in general entities but not predefined or character entities. Frankly, this confuses me. The XML spec (section 3.3.3) seems to outlaw '<' characters in attribute values completely, but expat definitely accepts them when they appear as character or predefined entities. * install.rb: added `--install-dir' command line option. * INSTALL: added description of `--install-dir' command line option. * Version 0.6.0 released. * examples/parseTestStream.rb: removed retry after parser errors. The parser usually gets too confused to continue. * examples/parseTestTree.rb: removed retry after parser errors. The parser usually gets too confused to continue. * nqxml/tokenizer.rb (nextText): completely rewritten to fix bug where '<' at beginning of text was substituted, then code thought a tag was starting there instead of passing through a literal '<'. (replaceAllRefsButParams): renamed from `replaceRefs'. * nqxml/entities.rb (NamedAttributes.attributesToXML): encode attribute values. (Tag.to_s): override simple source printing. 2001-04-06 Jim Menard * tests/writertester.rb: changed to reflect attribute tag handling. * tests/testresource.rb: @xml adds complex attribute value to a tag so we can test attribute value normalization. Text objects no longer pass XML source of text to constructors, except if they are CDATA. That's what happens when the tokenizer builds normal text and CDATA entities. * examples/reverseTags.rb: since text entities are no longer created with :source text when a document is parsed, we need to write the encoded text instead. * nqxml/entities.rb: Text entities no longer consider the :source attribute when comparing themselves. See next item for the reason. * nqxml/tokenizer.rb: renamed :generatedNegToken to :generatedEndTag. Rewritten to use a stack of input streams. (nextText): stopped giving XML source to Text entity constructor. This is because we can't keep track of the source during the complex process of general entity substitution (where new XML tags may be introduced in the middle of the text, for example). (skipSpaces): if spaces reach to end of current input stream, pop the stack and continue looking for spaces in the next input stream. (replaceParamRefs): created to handle parameter entity reference substitution within entity values. (nextEntityTag): added call to replaceParamRefs(). (normalizeAttributeValue): created. (initialize): line breaks in xml text normalized to #A. This allows us to use "\n" everywhere else in the tokenizer code, too. 2001-04-05 Jim Menard * tests/testresource.rb: changed class of @entityTag to GeneralEntityTag. * nqxml/entities.rb: added two new subclasses of EntityTag: GeneralEntityTag and ParameterEntityTag. (Comment.initialize): stopped interpreting contents of comments. * nqxml/tokenizer.rb (nextDeclSep): created to handle parameter entity references in DTD. (nextEntityTag): now recognizes parameter entity definitions and returnes one of GeneralEntityTag or ParameterEntityTag. (nextComment): disallow illegal comment ending string '--->'. When seen, a warning is printed to $stderr. (replaceEntityRefs): moved here from utils.rb because (a) only tokenizer should replace entity refs and (b) raising a parser error, newly added, requires a reference to the tokenizer anyway. (nextProcessingInstruction): do not allow space between ' * tests/tokenizertester.rb: fixed incorrect but harmless extra block parameters to Tokenizer.each() calls. * nqxml/utils.rb (replacePredefinedEntityRefs): separated out some of replaceEntityRefs() into this new method. (replaceCharacterRefs): separated out some of replaceEntityRefs() into this new method. (replaceEntityRefs): fixed regexp so '&&foo;' will correctly see '&foo;'. * nqxml/tokenizer.rb (nextEntityTag): ignore redefinition of entities, and print warning message. When reading an entity value, only parse character references ("&"), nothing else. * docs/README.sgml: reorganization, list of well-formedness rules not yet implemented, more. * nqxml/utils.rb (NQXML.replaceEntityRefs): made regexp for '&#xNN;' case-insensitive. * nqxml/tokenizer.rb (nextTagAttributes): added well-formedness check for '<' in attribute values. 2001-04-03 Jim Menard * nqxml/domparser.rb: added warning message. * release.rb: added check for up-to-date version numbers in a few doc files. * docs/sgml2text.rb: added section and example numbers, fixed sect3 and xrefs. No longer rely on hard-coded list of entity definitions; read them from SGML ENTITY tags. 2001-03-30 Jim Menard * docs/README.sgml: minor edits. * Version 0.5.1 released. * nqxml/entities.rb (Text.to_s): fixed Text.to_s to return text, not source. (Text.writeXMLTo): created to override the Entity version. Writes source if available; else writes encoded text. * tests/writertester.rb (test_writer): make sure writers properly encode special characters within text entity strings. * nqxml/utils.rb: created to hold methods useful for all but until recently held by the privileged few. * docs/README.sgml: updated to reflect the addition of epilogue to Document. * nqxml/treeparser.rb: three states define location in document: prolog, body, or epilogue. (handleNextEntity): split up into separate methods for easier readability. * nqxml/document.rb: added `:epilogue' attribute that holds comment, processing instruction, and white space elements that appear after the root node has closed (that is, at the end of the document). (addToProlog): fixed erroneous comment that said the tokenizer checked for certain entity types. It's really the tree parser. Removed useless commented-out code. (addToEpilogue): added. 2001-03-29 Jim Menard * docs/README.html: created so it can redirect from old URL at that location to the new README/t1.html. * docs/Makefile: reduced number of unnecessary warnings. Changed flags so HTML is generated as a directory with multiple HTML files instead of one big HTML file. Added `pdf' and `rtf' targets. * release.rb: as below: make sure we see the local version of info.rb instead of any previously installed one. Also modified to handle the HTML directory that contains the newly-expanded HTML version of the README file. * Version 0.5.0 released. * examples/*.rb: push '..' on to the front of $LOAD_PATH so we see the local version of NQXML instead of a previously installed one. * tests/test.rb: push '..' on to the front of $LOAD_PATH instead of at end. This way, even if there is an old version of NQXML installed we will run the one in this directory. * INSTALL: changed dir name `examples' to `tests'. * nqxml/document.rb: more argument error checking. Raise ArgumentError if certain arguments are nil. * nqxml/treeparser.rb (handleNextEntity): ignore whitespace outside of root node. * everywhere: changed entity constructors' argument orders and added default values to some paramaters to make manual creation of entities (therefore document objects) easier. * nqxml/entities.rb (NamedAttributes): factored entities out of Tag and ProcessingInstruction classes into a new NamedAttributes superclass. (NamedAttributes#attributesToXML): returns a string containing the attributes as XML. 2001-03-28 Jim Menard * nqxml/document.rb: created in order to separate the Document and Node classes from the tree parser. * tests/writertester.rb: Added test_string_arg and test_array_arg methods. * release.rb (WEB_DIR): requires and calls generateManifest. * generateManifest.rb: created. * nqxml/writer.rb: Changed from using `@io.print "foo"' to `@io << "foo"'. This way, IO, String, and Array objects may now be arguments to the Writer constructor. * tests/test.rb: split class Tester into a superclass and multiple subclasses that test the four major NQXML classes (tokenizer, streaming and tree parsers, and writer) and more. * tests/*: moved test.rb into this new subdirectory. * everywhere: renamed `DOMParser' class to `TreeParser'. To keep old scripts from breaking for now, there is still a `DOMParser' class that inherits from `TreeParser' and adds nothing else. `parseTestDOM.rb' renamed to `parseTestTree.rb'. * Version 0.4.2 released. 2001-03-23 Jim Menard * release.rb: generate documents and clean up before creating .tar.gz file. Munged ftp code. * Manifest: just keepin' it up to date. * docs/index.html: modified to refer to README.html (generated from README.sgml) instead of including that text. * docs/Makefile: created for making README and docs/README.html. * docs/sgml2text.rb: created for translating README.sgml into README. * docs/README.sgml: created DocBook version of README. * examples/data.xml: fixed malformed XML. * examples/parseTestDOM.rb: exception handling fixed. * examples/parseTestStream.rb: exception handling fixed. * examples/reverseTags.rb: exception handling fixed. * examples/reverseText.rb: exception handling fixed. Was using old error class name. * examples/printEntityClassNames.rb: created. 2001-03-22 Jim Menard * README: typed more words. Fixed examples. * docs/index.html: reformatted to fit my Web site and added colors to code comments. Added the words and fixes from README. * Version 0.4.1 released. 2001-03-19 Jim Menard * docs/index.html: added ``Writing XML'' section and fleshed out the list of examples. * README: added ``Writing XML'' section. 2001-03-13 Jim Menard * nqxml/streamingparser.rb (each): turned node stack `lastSeenTagNameStack' and a few other method-local variables into instance variables. * nqxml/domparser.rb (initialize): tured node stack `nodeStack' and a few other method-local variables into instance variables. 2001-03-12 Jim Menard * examples/write.rb: created. * test.rb: added tests for new Writer and WriterError classes. * nqxml/writer.rb: created; contains new Writer class. * nqxml/error.rb: added WriterError class. * release.rb: created. 2001-03-11 Jim Menard * nqxml/tokenizer.rb (peekMatches?): renamed from `peekMatches', and renamed all calls to this method. Created alias to old name. (eof?): renamed from `atEnd', and renamed all calls. Created alias to old name and to `eof' (with no question mark). * nqxml/parser.rb (eof?): renamed from `atEnd', and renamed all calls. Created alias to old name and to `eof' (with no question mark). 2001-03-09 Jim Menard * nqxml/streamingparser.rb (each): added some well-formedness checks. * nqxml/tokenizer.rb: added some well-formedness checks and consts NOT_SPACES_REGEX and PUBLIC_ID_LITERAL_REGEX. (skipSpaces): use NOT_SPACES_REGEX. (nextDoctype): added check for legal PUBLIC extern id literal chars. (nextPublicIdLiteral): added this method. * Version 0.4.0 released. * test.rb: added `expect_dom_error' and related tests. (expect_tokenizer_error): check exception line, col, and pos. (expect_streaming_error): ditto. (expect_dom_error): ditto. * README: updated examples. * docs/index.html: updated examples. * nqxml/tokenizer.rb (column): added column method. (initialize): made sure @pos set before checking argument type. * everywhere: renamed `SAXParser' class to `StreamingParser'. (everywhere): added tokenizer argument to all calls to ParserError constructors. * nqxml/error.rb (initialize): added `readStream' argument and readable :line, :column, and :pos attributes. 2001-03-08 Jim Menard * data.xml: created. * nqxml/entities.rb (Text#to_s): created this method to override Entity#to_s by returning the value of the @text attribute (instead of the original XML source, which may contain other entities like <). You can always call Text#source to get the original. * examples/reverseTags.rb: created. * examples/reverseText.rb: created. * doc.xml: Added a few tags, attributes, and values. * examples/dumpXML.rb: rewrote to use Ruby's feature of adding methods to already-existing classes. * nqxml/info.rb: Version string corrected (had "{VERSION_TWEAK}" instead of "#{VERSION_TWEAK}"). * Version 0.3.0 released. * examples/dumpXML.rb: created. * nqxml/tokenizer.rb (initialize): renamed `ioOrString' to `stringOrReadable'. (nextTagAttributes): created by factoring code from `nextTag' and `nextProcessingInstruction'. * nqxml/saxparser.rb (initialize): renamed `ioOrString' to `stringOrReadable'. * nqxml/parser.rb (initialize): renamed `ioOrString' to `stringOrReadable'. * nqxml/domparser.rb (initialize): renamed `ioOrString' to `stringOrReadable'. * README: renamed params in example code from `io_or_string' to `string_or_readable'. * docs/info.html: renamed params in example code from `io_or_string' to `string_or_readable'. General updates. 2001-03-07 Jim Menard * nqxml/tokenizer.rb added tag name to many error messages. (nextTag): added constraint that no attribute name may appear twice in the same start tag. (nextProcessingInstruction): ditto. (line): new method that returns current line number in XML. (nextDoctype): DOCTYPE's external id PUBLIC takes two args, SYSTEM takes one arg. I had it backwards. * test.rb (test_tokenizer_twin_attrs): added this test for the well-formedness constraint that no attribute name may appear twice in the same start tag or processing instruction. (expect_tokenizer_error): modified to take optional line number argument; if used, error must appear on the specified line. (expect_sax_error): ditto. (initReadOnlyVars): changed `PUBLIC' to `SYSTEM' in XML (related to bug where I had arguments to these reversed). (test_entity_replacement): renamed from `test_string_extension'. * examples/parserTestSAX.rb (testParser): report line number on error. * examples/parserTestDOM.rb (testParser): report line number on error. * nqxml/parser.rb (line): new method that passes request to tokenizer. * README: add link to Ruby's license. General updates. * nqxml/*.rb: added copyright and license notices. * nqxml/info.rb: created. * Version 0.2.2 released. * nqxml/tokenizer.rb (initialize): changed the way the type of the ioOrString argument to Tokenizer#initialize() is checked. Only Strings (or subclasses of String) or objects that have a 'read' method are allowed. Illegal argument exception text changed. From a bug report by David Alan Black. * test.rb (expect_tokenizer_error): created as a parallel to `expect_sax_error' test. * Version 0.2.1 released. * docs/index.html: fixed documentation bugs reported by David Alan Black. * README: fixed documentation bugs reported by David Alan Black. Deleted incorrect statement about entities within DOCTYPE tags being yielded before the DOCTYPE tag itself. * Version 0.2 released. * test.rb: changed tests because the SAX parser now yields entity instead of type, name, and data). Reorganized testing based on realization that tokenizer and SAX and DOM parsers return results in different orders. (initReadOnlyVars): added test XML for ELEMENT, ATTLIST, and NOTATION entities. * nqxml/saxparser.rb (each): stopped yielding type, name, and data (like expat does). Instead, we now yield the entity object itself. (each): added code to send ENTITY, ELEMENT, ATTLIST, and NOTATION entites. Note that these are sent before the DOCTYPE is sent because we have not yet seen the end of the DOCTYPE tag. * nqxml/entities.rb: created new classes Element, Attlist, and Notation. These are simple version that store all attributes and data as a single text blob. Future versions will implement these classes correctly. * nqxml/entities.rb: factored out :name attribute by creating new subclass of Entity called NamedEntity. 2001-03-06 Jim Menard * test.rb: wrote tests for DOM that helped catch a few bugs. * nqxml/parser.rb: moved SAX and DOM parsers into separate files saxparser.rb and domparser.rb. You now have to require 'nqxml/saxparser' and/or require 'nqxml/domparser', where before you could just require 'nqxml/parser'. This saves having to load an unwanted parser. It also cleans up the code. * examples: created directory and moved parseTest.rb there. * nqxml/tokenizer.rb: started writing code to handle the ELEMENT, ATTLIST, and NOTATION tags, processing instructions, and comments within DOCTYPE tags. * install.rb: fixed bad call of Find.find() (reported by TAKAHASHI 'Maki' Masayoshi). * nqxml/entities.rb: removed unused attributes :isTag, :isComment, etc. from Entity class and subclasses. * ./: added Manifest and Credits files. 2001-03-05 Jim Menard * test.rb: wrote tests for DOCTYPE and ENTITY tags and for entity replacement within text. * nqxml/tokenizer.rb: added code to replace entities defined with the tag when they are seen in text. * nqxml/tokenizer.rb: started writing code to handle the DOCTYPE tag and ENTITY tags within DOCTYPE. * nqxml/entities.rb: added Entity subclasses EntityTag () and Doctype () * nqxml/tokenizer.rb: renamed `Tokenizer#nextAttributeValue' to `Tokenizer#nextQuotedLiteral'. * nqxml/parser.rb: renamed Parser attribute :token to :entity. * nqxml/entities.rb: renamed Token class to Entity. * nqxml/entities.rb: created from classes formerly defined in nqxml/tokenizer.rb. * Version 0.1 released. Initial release.