emelie/tidy-html5
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

Issue #130 - initial MathML support

Browse source
This commit is contained in:
Geoff McLane 2015-02-05 12:21:08 +01:00
parent 5f470763c5
commit 1be5ccbb63
6 changed files with 45 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
include/tidyenum.h
View file

@ -384,6 +384,7 @@ typedef enum
TidyTag_LINK, /**< LINK */ TidyTag_LINK, /**< LINK */
TidyTag_LISTING, /**< LISTING */ TidyTag_LISTING, /**< LISTING */
TidyTag_MAP, /**< MAP */ TidyTag_MAP, /**< MAP */
TidyTag_MATHML, /**< MATH (HTML5) [i_a]2 MathML embedded in [X]HTML */
TidyTag_MARQUEE, /**< MARQUEE */ TidyTag_MARQUEE, /**< MARQUEE */
TidyTag_MENU, /**< MENU */ TidyTag_MENU, /**< MENU */
TidyTag_META, /**< META */ TidyTag_META, /**< META */

26
src/attrdict.c
View file

@ -7748,6 +7748,32 @@ const AttrVersion TY_(W3CAttrsFor_MAP)[] =
{ TidyAttr_UNKNOWN, 0 }, { TidyAttr_UNKNOWN, 0 },
}; };
const AttrVersion TY_(W3CAttrsFor_MATHML)[] = /* [i_a]2 */
{
{ TidyAttr_ALIGN, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_CLASS, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_DIR, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_ID, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_HEIGHT, xxxx|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50 },
{ TidyAttr_LANG, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|xxxx|xxxx|HT50|XH50 },
{ TidyAttr_OnCLICK, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnDBLCLICK, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnKEYDOWN, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnKEYPRESS, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnKEYUP, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnMOUSEDOWN, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnMOUSEMOVE, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnMOUSEOUT, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnMOUSEOVER, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_OnMOUSEUP, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_STYLE, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_TITLE, xxxx|xxxx|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_WIDTH, xxxx|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50 },
{ TidyAttr_XML_LANG, xxxx|xxxx|xxxx|xxxx|X10T|xxxx|xxxx|X10F|xxxx|xxxx|X10S|XH11|xxxx|HT50|XH50 },
{ TidyAttr_XMLNS, xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|XH11|xxxx|HT50|XH50 },
{ TidyAttr_UNKNOWN, 0 },
};
const AttrVersion TY_(W3CAttrsFor_MARK)[] = const AttrVersion TY_(W3CAttrsFor_MARK)[] =
{ {
{ TidyAttr_ACCESSKEY, xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|HT50|XH50 }, { TidyAttr_ACCESSKEY, xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|HT50|XH50 },

1
src/attrdict.h
View file

@ -72,6 +72,7 @@ extern const AttrVersion TY_(W3CAttrsFor_LI)[];
extern const AttrVersion TY_(W3CAttrsFor_LINK)[]; extern const AttrVersion TY_(W3CAttrsFor_LINK)[];
extern const AttrVersion TY_(W3CAttrsFor_LISTING)[]; extern const AttrVersion TY_(W3CAttrsFor_LISTING)[];
extern const AttrVersion TY_(W3CAttrsFor_MAP)[]; extern const AttrVersion TY_(W3CAttrsFor_MAP)[];
extern const AttrVersion TY_(W3CAttrsFor_MATHML)[]; /* [i_a]2 */
extern const AttrVersion TY_(W3CAttrsFor_MENU)[]; extern const AttrVersion TY_(W3CAttrsFor_MENU)[];
extern const AttrVersion TY_(W3CAttrsFor_META)[]; extern const AttrVersion TY_(W3CAttrsFor_META)[];
extern const AttrVersion TY_(W3CAttrsFor_NEXTID)[]; extern const AttrVersion TY_(W3CAttrsFor_NEXTID)[];

7
src/lexer.c
View file

@ -2525,7 +2525,7 @@ static Node* GetTokenFromStream( TidyDocImpl* doc, GetTokenMode mode )
c = ParseTagName( doc ); c = ParseTagName( doc );
isempty = no; isempty = no;
attributes = NULL; attributes = NULL;
lexer->token = TagToken( doc, (isempty ? StartEndTag : StartTag) ); lexer->token = TagToken( doc, StartTag ); /* [i_a]2 'isempty' is always false, thanks to code 2 lines above */
/* parse attributes, consuming closing ">" */ /* parse attributes, consuming closing ">" */
if (c != '>') if (c != '>')
@ -2562,7 +2562,10 @@ static Node* GetTokenFromStream( TidyDocImpl* doc, GetTokenMode mode )
lexer->state = LEX_CONTENT; lexer->state = LEX_CONTENT;
if (lexer->token->tag == NULL) if (lexer->token->tag == NULL)
TY_(ReportFatal)( doc, NULL, lexer->token, UNKNOWN_ELEMENT ); {
if (mode != OtherNamespace) /* [i_a]2 only issue warning if NOT 'OtherNamespace', and tag null */
TY_(ReportFatal)( doc, NULL, lexer->token, UNKNOWN_ELEMENT );
}
else if ( !cfgBool(doc, TidyXmlTags) ) else if ( !cfgBool(doc, TidyXmlTags) )
{ {
Node* curr = lexer->token; Node* curr = lexer->token;

10
src/parser.c
View file

@ -734,6 +734,10 @@ static Bool InsertMisc(Node *element, Node *node)
static void ParseTag( TidyDocImpl* doc, Node *node, GetTokenMode mode ) static void ParseTag( TidyDocImpl* doc, Node *node, GetTokenMode mode )
{ {
Lexer* lexer = doc->lexer; Lexer* lexer = doc->lexer;
if (node->tag == NULL) /* [i_a]2 prevent crash for active content (php, asp) docs */
return;
/* /*
Fix by GLP 2000-12-21. Need to reset insertspace if this Fix by GLP 2000-12-21. Need to reset insertspace if this
is both a non-inline and empty tag (base, link, meta, isindex, hr, area). is both a non-inline and empty tag (base, link, meta, isindex, hr, area).
@ -753,6 +757,8 @@ static void ParseTag( TidyDocImpl* doc, Node *node, GetTokenMode mode )
if (node->type == StartEndTag) if (node->type == StartEndTag)
return; return;
lexer->parent = node; /* [i_a]2 added this - not sure why - CHECKME: */
(*node->tag->parser)( doc, node, mode ); (*node->tag->parser)( doc, node, mode );
} }
@ -3908,7 +3914,7 @@ void TY_(ParseBody)(TidyDocImpl* doc, Node *body, GetTokenMode mode)
if (TY_(nodeIsElement)(node)) if (TY_(nodeIsElement)(node))
{ {
if ( TY_(nodeHasCM)(node, CM_INLINE) ) if ( TY_(nodeHasCM)(node, CM_INLINE) && !TY_(nodeHasCM)(node, CM_MIXED) ) /* [i_a]2 add CM_MIXED */
{ {
/* HTML4 strict doesn't allow inline content here */ /* HTML4 strict doesn't allow inline content here */
/* but HTML2 does allow img elements as children of body */ /* but HTML2 does allow img elements as children of body */
@ -4483,7 +4489,7 @@ static void AttributeChecks(TidyDocImpl* doc, Node* node)
if (TY_(nodeIsElement)(node)) if (TY_(nodeIsElement)(node))
{ {
if (node->tag->chkattrs) if (node->tag && node->tag->chkattrs) /* [i_a]2 fix crash after adding SVG support with alt/unknown tag subtree insertion there */
node->tag->chkattrs(doc, node); node->tag->chkattrs(doc, node);
else else
TY_(CheckAttributes)(doc, node); TY_(CheckAttributes)(doc, node);

4
src/tags.c
View file

@ -77,6 +77,7 @@ static CheckAttribs CheckHTML;
#define VERS_ELEM_LINK (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50) #define VERS_ELEM_LINK (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50)
#define VERS_ELEM_LISTING (HT20|HT32|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx) #define VERS_ELEM_LISTING (HT20|HT32|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx)
#define VERS_ELEM_MAP (xxxx|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50) #define VERS_ELEM_MAP (xxxx|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|xxxx|HT50|XH50)
#define VERS_ELEM_MATHML (xxxx|xxxx|xxxx|H41T|X10T|xxxx|H41F|X10F|xxxx|H41S|X10S|XH11|xxxx|HT50|XH50) /* [i_a]2 */
#define VERS_ELEM_MENU (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|xxxx|xxxx|xxxx|xxxx|xxxx|HT50|XH50) #define VERS_ELEM_MENU (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|xxxx|xxxx|xxxx|xxxx|xxxx|HT50|XH50)
#define VERS_ELEM_META (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50) #define VERS_ELEM_META (HT20|HT32|H40T|H41T|X10T|H40F|H41F|X10F|H40S|H41S|X10S|XH11|XB10|HT50|XH50)
#define VERS_ELEM_NEXTID (HT20|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx) #define VERS_ELEM_NEXTID (HT20|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx|xxxx)
@ -218,7 +219,8 @@ static const Dict tag_defs[] =
{ TidyTag_LINK, "link", VERS_ELEM_LINK, &TY_(W3CAttrsFor_LINK)[0], (CM_HEAD|CM_BLOCK|CM_EMPTY), TY_(ParseEmpty), CheckLINK }, { TidyTag_LINK, "link", VERS_ELEM_LINK, &TY_(W3CAttrsFor_LINK)[0], (CM_HEAD|CM_BLOCK|CM_EMPTY), TY_(ParseEmpty), CheckLINK },
{ TidyTag_LISTING, "listing", VERS_ELEM_LISTING, &TY_(W3CAttrsFor_LISTING)[0], (CM_BLOCK|CM_OBSOLETE), TY_(ParsePre), NULL }, { TidyTag_LISTING, "listing", VERS_ELEM_LISTING, &TY_(W3CAttrsFor_LISTING)[0], (CM_BLOCK|CM_OBSOLETE), TY_(ParsePre), NULL },
{ TidyTag_MAP, "map", VERS_ELEM_MAP, &TY_(W3CAttrsFor_MAP)[0], (CM_INLINE), TY_(ParseBlock), NULL }, { TidyTag_MAP, "map", VERS_ELEM_MAP, &TY_(W3CAttrsFor_MAP)[0], (CM_INLINE), TY_(ParseBlock), NULL },
// { TidyTag_MENU, "menu", VERS_ELEM_MENU, &TY_(W3CAttrsFor_MENU)[0], (CM_BLOCK|CM_OBSOLETE), TY_(ParseList), NULL }, { TidyTag_MATHML, "math", VERS_ELEM_MATHML, &TY_(W3CAttrsFor_MATHML)[0], (CM_INLINE|CM_BLOCK|CM_MIXED), TY_(ParseNamespace),NULL }, /* [i_a]2 */
/* { TidyTag_MENU, "menu", VERS_ELEM_MENU, &TY_(W3CAttrsFor_MENU)[0], (CM_BLOCK|CM_OBSOLETE), TY_(ParseList), NULL }, */
{ TidyTag_META, "meta", VERS_ELEM_META, &TY_(W3CAttrsFor_META)[0], (CM_HEAD|CM_BLOCK|CM_EMPTY), TY_(ParseEmpty), NULL }, { TidyTag_META, "meta", VERS_ELEM_META, &TY_(W3CAttrsFor_META)[0], (CM_HEAD|CM_BLOCK|CM_EMPTY), TY_(ParseEmpty), NULL },
{ TidyTag_NOFRAMES, "noframes", VERS_ELEM_NOFRAMES, &TY_(W3CAttrsFor_NOFRAMES)[0], (CM_BLOCK|CM_FRAMES), TY_(ParseNoFrames), NULL }, { TidyTag_NOFRAMES, "noframes", VERS_ELEM_NOFRAMES, &TY_(W3CAttrsFor_NOFRAMES)[0], (CM_BLOCK|CM_FRAMES), TY_(ParseNoFrames), NULL },
{ TidyTag_NOSCRIPT, "noscript", VERS_ELEM_NOSCRIPT, &TY_(W3CAttrsFor_NOSCRIPT)[0], (CM_HEAD|CM_BLOCK|CM_INLINE|CM_MIXED), TY_(ParseBlock), NULL }, { TidyTag_NOSCRIPT, "noscript", VERS_ELEM_NOSCRIPT, &TY_(W3CAttrsFor_NOSCRIPT)[0], (CM_HEAD|CM_BLOCK|CM_INLINE|CM_MIXED), TY_(ParseBlock), NULL },