* Various patches, see ChangeLog

ChangeLog

@@ -1,3 +1,13 @@
+* FIXED: Proper parsing of <!NOTATION ...> declaration.  Also fixed
+  dtd_property(DTD, notations(Notations)) and modified
+  dtd_property(DTD, notation(N, Decl)).
+
+* ADDED: <!ENTITY #DEFAULT ...> (Richard O'Keefe).
+
+* FIXED: Ignore newline after entity
+
+* FIXED: Allow omitted ; expanding entities in CDATA attributes
+
 * FIXED: Possible crash with SHORTREF endding in &#RE.
 
 * FIXED: avoid crash on illegal syntax in DTD omited-tag declaration

Test/bar.sgml

@@ -0,0 +1,8 @@
+<!DOCTYPE bar [
+ <!ELEMENT bar - O EMPTY>
+ <!ATTLIST bar a CDATA #REQUIRED b NAME #REQUIRED c NUMBER #REQUIRED
+               d CDATA #REQUIRED>
+]>
+
+<bar a="Major Mynah" b="&#65;&#66;&#67;" c="&#49;&#50;" d="foo
+bar     ugh">

Test/ce.sgml

@@ -0,0 +1,7 @@
+<!DOCTYPE test [
+  <!ELEMENT test - - (#PCDATA)>
+]>
+
+<test>
+Test for handling character entities: &#65; &#RS;
+</test>

Test/cmt.sgml

@@ -0,0 +1,13 @@
+<!DOCTYPE a
+ [ <!ENTITY %attrs
+   "id	ID #implied -- document-wide unique id --
+    class CDATA #IMPLIED -- comment --">
+
+   <!ELEMENT a - - (#PCDATA) -- foo-bar -->
+   <!ATTLIST a %attrs;>
+ ]
+>
+
+<!-- x -->
+
+<a></a>

Test/conref.sgml

@@ -0,0 +1,8 @@
+<!DOCTYPE con [
+ <!ELEMENT con - - (foo,bar)>
+ <!ATTLIST con missing (missing) #CONREF>
+ <!ELEMENT (foo|bar) - O (#PCDATA)>
+]>
+
+<con missing>
+

Test/defent.sgml

@@ -0,0 +1,10 @@
+<!DOCTYPE def [
+  <!ENTITY #DEFAULT SDATA "?">
+  <!ELEMENT def - - RCDATA>
+]>
+
+<def>&#65
+&#66 &#67
+&b
+&#66
+&c</def>

Test/foo.sgml

@@ -0,0 +1,7 @@
+<!DOCTYPE foo [ 
+ <!ELEMENT foo O O (name+)>
+ <!ELEMENT name - O (#PCDATA)>
+ <!ATTLIST name title cdata #implied sex cdata #implied>
+]>
+<name title=Dr sex=M>Neddie Seagoon
+<name title=Miss sex=F>Minnie Bannister

Test/i.sgml

@@ -0,0 +1,12 @@
+<!DOCTYPE test [
+  <!ENTITY QS "<q>">
+  <!ENTITY QE ENDTAG "q">
+  <!ENTITY Q2 CDATA "<q>">
+  <!ENTITY S SDATA "special">
+  <!ELEMENT test - - (#PCDATA|q)*>
+  <!ELEMENT q    - - (#PCDATA)>
+]>
+
+<test>
+&Q2;! Hello &QS;you there&QE;, I'm &S;
+</test>

Test/not.sgml

@@ -0,0 +1,13 @@
+<!DOCTYPE not [
+  <!ELEMENT not - - (Formula*)>
+  <!NOTATION TeX SYSTEM "">
+  <!NOTATION Eqn SYSTEM "">
+  <!ELEMENT formula - O CDATA>
+  <!ATTLIST formula notation NOTATION (TeX|Eqn) #CURRENT>
+]>
+
+<not>
+<formula notation=TeX>
+$$E = MC^2$$
+</formula>
+</not>

Test/ok/bar.ok

@@ -0,0 +1 @@
+[element(bar, [a='Major Mynah', b=abc, c='12', d='foo bar     ugh'], [])].

Test/ok/ce.ok

@@ -0,0 +1 @@
+[element(test, [], ['Test for handling character entities: A \n'])].

Test/ok/cmt.ok

@@ -0,0 +1 @@
+[element(a, [], [])].

Test/ok/conref.ok

@@ -0,0 +1 @@
+[element(con, [missing=missing], [])].

Test/ok/defent.ok

@@ -0,0 +1 @@
+[element(def, [], ['AB C', sdata(?), 'B', sdata(?)])].

Test/ok/foo.ok

@@ -0,0 +1 @@
+[element(foo, [], [element(name, [title='Dr', sex='M'], ['Neddie Seagoon']), element(name, [title='Miss', sex='F'], ['Minnie Bannister'])])].

Test/ok/i.ok

@@ -0,0 +1 @@
+[element(test, [], ['<q>! Hello ', element(q, [], ['you there']), ', I\'m ', sdata(special)])].

Test/ok/not.ok

@@ -0,0 +1 @@
+[element(not, [], [element(formula, [notation=tex], ['$$E = MC^2$$'])])].

Test/ok/omit1.ok

@@ -0,0 +1 @@
+[element(bar, [], [element(foo, [], [element(a, [], [element(b, [], [x]), element(c, [], [y])]), element(a, [], [element(b, [], [x]), element(c, [], [y])]), element(a, [], [element(b, [], [x]), element(c, [], [y])]), element(a, [], [element(b, [], [x]), element(c, [], [y])])])])].

Test/ok/per.ok

@@ -0,0 +1 @@
+[element(per, [], [element(jim, [], [snark])])].

Test/ok/rcdata.ok

@@ -0,0 +1 @@
+[element(doc, [], [element(t, [], ['This is rcdata Hello & World Hello']), element(a, [], ['This is an a'])])].

Test/ok/sdata.ok

@@ -0,0 +1 @@
+[element(text, [], ['Van Alpha tot ', sdata('\\Omega')])].

Test/ok/shortval.ok

@@ -0,0 +1 @@
+[element(num, [val='2'], [])].

Test/ok/sr.ok

@@ -0,0 +1 @@
+[element(test, [], [element(p, [], ['Peter said: ', element(q, [], ['He, this is a nice program']), '.']), element(p, [], ['Bob said: ', element(q, [], ['Yes, it is'])])])].

Test/ok/sr2.ok

@@ -0,0 +1 @@
+[element(test, [], [element(t, [], [element(name, [], [n]), element(arg, [], [a])])])].

Test/ok/ugh.ok

@@ -0,0 +1 @@
+[element(ugh, [], [element(extension, [missing=missing, exchange='479'], []), element(extension, [exchange='479'], ['8494']), element(extension, [exchange='555'], ['1234']), element(extension, [exchange='03-555'], ['1234'])])].

Test/omit1.sgml

@@ -0,0 +1,14 @@
+<!DOCTYPE bar [
+  <!ELEMENT bar - - (foo)>
+  <!ELEMENT foo O O (a+)>
+  <!ELEMENT a   O O (b,c)>
+  <!ELEMENT b   - O (#PCDATA)>
+  <!ELEMENT c   - O (#PCDATA)>
+]>
+
+<bar>
+<b>x<c>y
+<b>x<c>y
+<b>x<c>y
+<b>x<c>y
+</bar>

Test/per.sgml

@@ -0,0 +1,9 @@
+<!DOCTYPE per [
+  <!ENTITY % fred "jim">
+  <!ELEMENT per - - (%fred;)>
+  <!ELEMENT %fred - - (#PCDATA)>
+  <!ATTLIST %fred tom (dick|harry) #IMPLIED>
+  <!USEMAP #EMPTY %fred>
+]>
+
+<per><jim>snark</jim></per>

Test/rcdata.sgml

@@ -0,0 +1,19 @@
+<!DOCTYPE doc
+ [ <!ELEMENT doc - - (t|a)*>
+   <!ELEMENT a   - - (#PCDATA)>
+   <!ELEMENT t - - rcdata>
+   <!ENTITY  hello "Hello">
+   <!ENTITY  world "World &hello;">
+ ]
+>
+
+<!-- coment -->
+
+<doc>
+<t>
+This is rcdata &hello; & &world
+</t>
+
+<a>This is an a</a>
+
+</doc>

Test/sdata.sgml

@@ -0,0 +1,8 @@
+<!DOCTYPE text [
+  <!ENTITY Omega SDATA "\Omega">
+  <!ELEMENT text - - (#PCDATA)>
+]>
+
+<text>
+Van Alpha tot &Omega;
+</text>

Test/shortval.sgml

@@ -0,0 +1,6 @@
+<!DOCTYPE num [
+  <!ELEMENT num - O EMPTY>
+  <!ATTLIST num val (1|2|3) #REQUIRED>
+]>
+
+<num 2>

Test/sr.sgml

@@ -0,0 +1,20 @@
+<!DOCTYPE test [
+  <!ELEMENT test - - p*>
+  <!ELEMENT p    o o (#PCDATA|q)*>
+  <!ELEMENT q    - - (#PCDATA)>
+
+  <!ENTITY   beg..q "<q><!USEMAP in..q>">
+  <!ENTITY   end..q ENDTAG "q">
+  <!ENTITY   beg..p STARTTAG "p">
+
+  <!SHORTREF in..p  '"' beg..q
+		    '&#RS;B&#RE' beg..p>
+  <!SHORTREF in..q  '"' end..q>
+  <!USEMAP in..p p>
+]>
+
+<test>
+Peter said: "He, this is a nice program".
+
+Bob said: "Yes, it is"
+</test>

Test/sr2.sgml

@@ -0,0 +1,25 @@
+<!DOCTYPE test
+ [ <!ELEMENT test - - (t*)>
+   <!ELEMENT t - - ((name|op),arg)>
+   <!ELEMENT name o o (#PCDATA)>
+   <!ELEMENT op   - - (#PCDATA)>
+   <!ELEMENT arg  - - (#PCDATA)>
+   <!ELEMENT desc  o o (#PCDATA)>
+
+   <!ENTITY end-t ENDTAG "t">
+   <!ENTITY beg-arg STARTTAG "arg">
+   <!ENTITY end-arg ENDTAG "arg">
+
+   <!SHORTREF in.t
+	'&#RE'  end-t
+        '('     beg-arg>
+   <!SHORTREF in.arg
+	')'	end-arg>
+
+   <!USEMAP in.t t>
+   <!USEMAP in.arg arg>
+ ]>
+
+<test>
+<t>n(a)
+</test>

Test/test.pl

@@ -0,0 +1,103 @@
+/*  $Id$
+
+    Part of SWI-Prolog SGML/XML parser
+
+    Author:  Jan Wielemaker
+    E-mail:  [email protected]
+    WWW:     http://www.swi.psy.uva.nl/projects/SWI-Prolog/
+    Copying: LGPL-2.  See the file COPYING or http://www.gnu.org
+
+    Copyright (C) 1990-2000 SWI, University of Amsterdam. All rights reserved.
+*/
+
+:- module(sgml_test,
+	  [ test/1,			% +File
+	    testdir/1,			% +Dir
+	    pass/1,			% +File
+	    show/1,			% +File
+	    test/0
+	  ]).
+
+test :-
+	testdir(.).
+
+testdir(Dir) :-
+	atom_concat(Dir, '/*', Pattern),
+	expand_file_name(Pattern, Files),
+	checklist(dotest, Files).
+
+dotest(File) :-
+	file_name_extension(_, Ext, File),
+	memberchk(Ext, [sgml, xml, html]), !,
+	test(File).
+dotest(_).
+
+test(File) :-
+	format('Test ~w ... ', [File]),
+	flush_output,
+	load_file(File, Term),
+	ground(Term),			% make sure
+	okfile(File, OkFile),
+	(   exists_file(OkFile)
+	->  load_prolog_file(OkFile, TermOk),
+	    (	compare_dom(Term, TermOk)
+	    ->	format('ok~n')
+	    ;	format('WRONG~n')
+	    )
+	;   format('Loaded, no validating data~n'),
+	    pp(Term)
+	).
+
+show(File) :-
+	load_file(File, Term),
+	pp(Term).
+
+pass(File) :-
+	load_file(File, Term),
+	okfile(File, OkFile),
+	open(OkFile, write, Fd),
+	format(Fd, '~q.~n', [Term]),
+	close(Fd).
+
+load_file(File, Term) :-
+	load_pred(Ext, Pred),
+	file_name_extension(_, Ext, File), !,
+	call(Pred, File, Term).
+load_file(Base, Term) :-
+	load_pred(Ext, Pred),
+	file_name_extension(Base, Ext, File),
+	exists_file(File), !,
+	call(Pred, File, Term).
+
+
+load_pred(sgml,	load_sgml_file).
+load_pred(xml,	load_xml_file).
+load_pred(html,	load_html_file).
+
+okfile(File, OkFile) :-
+	file_name_extension(Base, _, File),
+	file_directory_name(Base, Dir),
+	concat_atom([Dir, '/ok/', Base, '.ok'], OkFile).
+
+load_prolog_file(File, Term) :-
+	open(File, read, Fd),
+	read(Fd, Term),
+	close(Fd).
+
+compare_dom([], []) :- !.
+compare_dom([H1|T1], [H2|T2]) :- !,
+	compare_dom(H1, H2),
+	compare_dom(T1, T2).
+compare_dom(X, X) :- !.
+compare_dom(element(Name, A1, Content1),
+	    element(Name, A2, Content2)) :-
+	compare_attributes(A1, A2),
+	compare_dom(Content1, Content2).
+
+compare_attributes(A1, A2) :-
+	sort(A1, L1),
+	sort(A2, L2),
+	L1 == L2.
+
+
+

Test/ugh.sgml

@@ -0,0 +1,9 @@
+<!DOCTYPE ugh [
+ <!ELEMENT ugh O O (extension+)>
+ <!ELEMENT extension - O (#PCDATA)>
+ <!ATTLIST extension exchange NUTOKEN 479 missing (missing) #CONREF>
+]>
+<extension missing>
+<extension>8494
+<extension exchange=555>1234
+<extension exchange=03-555>1234

doc/sgml2pl.sgml

@@ -607,8 +607,9 @@ unbound.
 </argdef>
     <termitem>notations(ListOfNotations)
 Returns a list holding the names of all <c/NOTATION/ declarations. 
-    <termitem>notation(Name, File)
-Yields the declared file for from a <c/NOTATION/ declaration. 
+    <termitem>notation(Name, Decl)
+Unify <aref/Decl/ with a list if <term>system(+File)</term> and/or
+<term>public(+PublicId)</term>.
 </argdef>
 </definitions>
 

dtd.h

@@ -282,7 +282,9 @@ typedef struct _dtd_entity
 
 typedef struct _dtd_notation
 { dtd_symbol *name;			/* name of the notation */
-  ichar *file;				/* file with info */
+  entity_type type;			/* ET_{PUBLIC|SYSTEM} */
+  ichar *public;			/* public id */
+  ichar *system;			/* file with info */
   struct _dtd_notation *next;		/* list-link */
 } dtd_notation;
 
@@ -401,6 +403,7 @@ typedef struct _dtd
   dtd_symbol_table     *symbols;	/* symbol-table */
   dtd_entity           *pentities;	/* defined parameter entities */
   dtd_entity	       *entities;	/* defined entities */
+  dtd_entity	       *default_entity;	/* default-entity (if any) */
   dtd_notation	       *notations;	/* Declared notations */
   dtd_shortref	       *shortrefs;	/* SHORTREF declarations */
   dtd_element          *elements;	/* defined elements */