experiments/clixon
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

* XML parser conformance to W3 spec

Browse source 
* Names lexically correct (NCName)
  * Syntactically Correct handling of '<?' (processing instructions) and '<?xml' (XML declaration)
  * XML prolog syntax for 'well-formed' XML
  * <!DOCTYPE (ie DTD) is not supported.
This commit is contained in:
Olof hagsand 2018-11-18 13:22:08 +01:00
parent 9bd0dc42c6
commit 9c57902b96
9 changed files with 280 additions and 69 deletions
Download patch file Download diff file Expand all files Collapse all files

61
lib/src/clixon_xml_parse.l
View file

@ -34,6 +34,8 @@
* XML parser
* @see https://www.w3.org/TR/2008/REC-xml-20081126
* https://www.w3.org/TR/2009/REC-xml-names-20091208
*
*/
%{
@ -72,8 +74,22 @@ int clixon_xml_parsewrap(void)
return 1;
}
/*
* From https://www.w3.org/TR/2008/REC-xml-20081126:
* [4]* NameStartChar ::= ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] ...
* [4a] NameChar ::= NameStartChar | "-" | "." | [0-9] | #xB7|
* [5] Name ::= NameStartChar (NameChar)*
* NOTE: From https://www.w3.org/TR/2009/REC-xml-names-20091208:
* [4] NCName ::= Name - (Char* ':' Char*) An XML Name, minus the ":"
* --> namestart and name below is NCNAME
*/
%}
namestart [A-Z_a-z]
namechar [A-Z_a-z\-\.0-9]
ncname {namestart}{namechar}*
%x START
%s STATEA
%s AMPERSAND
@ -81,36 +97,42 @@ int clixon_xml_parsewrap(void)
%s CMNT
%s STR
%s TEXTDECL
%s PIDECL
%s PIDECL2
%s STRDQ
%s STRSQ
%%
<START>[0-9A-Za-z_\-]+ { clixon_xml_parselval.string = strdup(yytext);
<START,TEXTDECL>[ \t] ;
<START,STATEA,CMNT,TEXTDECL>\n { _YA->ya_linenum++; }
<START>{ncname} { clixon_xml_parselval.string = strdup(yytext);
return NAME; /* rather be catch-all */
}
<START>[ \t]+ ;
<START>\: return *clixon_xml_parsetext;
<START>\n { _YA->ya_linenum++;}
<START>"<?xml" { BEGIN(TEXTDECL); return BTEXT;}
<START><<EOF>> { return MY_EOF; }
<START>"<?xml" { BEGIN(TEXTDECL); return BXMLDCL;}
<START>"<?" { BEGIN(PIDECL); return BQMARK;}
<START>"/>" { BEGIN(STATEA); return ESLASH; }
<START>"<!--" { BEGIN(CMNT); return BCOMMENT; }
<START>"</" return BSLASH;
<START>[/=] return *clixon_xml_parsetext;
<START>\< return *clixon_xml_parsetext;
<START>\> { BEGIN(STATEA); return *clixon_xml_parsetext; }
<START>\" { _YA->ya_lex_state=START;BEGIN(STRDQ); return *clixon_xml_parsetext; }
<START>\' { _YA->ya_lex_state=START;BEGIN(STRSQ); return *clixon_xml_parsetext; }
<START>. { clixon_xml_parselval.string = yytext; return CHARDATA; /*XXX:optimize*/ }
<START>. { clixon_xml_parselval.string = yytext; return CHARDATA; /* optimize? */}
<STATEA>"</" { BEGIN(START); return BSLASH; }
<STATEA><<EOF>> { return MY_EOF; }
<STATEA>"<!--" { BEGIN(CMNT); return BCOMMENT; }
<STATEA>"<![CDATA[" { BEGIN(CDATA); _YA->ya_lex_state = STATEA; clixon_xml_parselval.string = yytext; return CHARDATA;}
<STATEA>"<?" { BEGIN(PIDECL); return BQMARK; }
<STATEA>\< { BEGIN(START); return *clixon_xml_parsetext; }
<STATEA>& { _YA->ya_lex_state =STATEA;BEGIN(AMPERSAND);}
<STATEA>\n { clixon_xml_parselval.string = yytext;_YA->ya_linenum++; return (CHARDATA);}
<STATEA>. { clixon_xml_parselval.string = yytext; return CHARDATA; /*XXX:optimize*/}
<STATEA>. { clixon_xml_parselval.string = yytext; return CHARDATA; }
/* @see xml_chardata_encode */
<AMPERSAND>"amp;" { BEGIN(_YA->ya_lex_state); clixon_xml_parselval.string = "&"; return CHARDATA;}
@ -124,21 +146,30 @@ int clixon_xml_parsewrap(void)
<CDATA>"]]>" { BEGIN(_YA->ya_lex_state); clixon_xml_parselval.string = yytext; return CHARDATA;}
<CMNT>"-->" { BEGIN(START); return ECOMMENT; }
<CMNT>\n _YA->ya_linenum++;
<CMNT>.
<TEXTDECL>encoding return ENC;
<TEXTDECL>version return VER;
<TEXTDECL>"=" return *clixon_xml_parsetext;
<TEXTDECL>"?>" { BEGIN(START);return ETEXT;}
<TEXTDECL>standalone return SD;
<TEXTDECL>"=" { return *clixon_xml_parsetext; }
<TEXTDECL>"?>" { BEGIN(START);return EQMARK;}
<TEXTDECL>\" { _YA->ya_lex_state =TEXTDECL;BEGIN(STRDQ); return *clixon_xml_parsetext; }
<TEXTDECL>\' { _YA->ya_lex_state =TEXTDECL;BEGIN(STRSQ); return *clixon_xml_parsetext; }
<TEXTDECL>. { clixon_xml_parselval.string = yytext; return CHARDATA; /* optimize? */}
<STRDQ>1\.[0-9]+ { clixon_xml_parselval.string = strdup(yytext); return CHARDATA; }
<STRDQ>[^\"]+ { clixon_xml_parselval.string = strdup(yytext); return CHARDATA; }
<PIDECL>{ncname} { clixon_xml_parselval.string = strdup(yytext);
return NAME; /* rather be catch-all */
}
<PIDECL>[ \t] { BEGIN(PIDECL2);}
<PIDECL>. { clixon_xml_parselval.string = yytext; return CHARDATA; /* optimize? */}
<PIDECL2>"?>" { BEGIN(START);return EQMARK;}
<PIDECL2>[^{?>}]+ { clixon_xml_parselval.string = strdup(yytext); return STRING; }
<STRDQ>1\.[0-9]+ { clixon_xml_parselval.string = strdup(yytext); return STRING; }
<STRDQ>[^\"]+ { clixon_xml_parselval.string = strdup(yytext); return STRING; }
<STRDQ>\" { BEGIN(_YA->ya_lex_state); return *clixon_xml_parsetext; }
<STRSQ>1\.[0-9]+ { clixon_xml_parselval.string = strdup(yytext); return CHARDATA; }
<STRSQ>[^\']+ { clixon_xml_parselval.string = strdup(yytext); return CHARDATA; }
<STRSQ>1\.[0-9]+ { clixon_xml_parselval.string = strdup(yytext); return STRING; }
<STRSQ>[^\']+ { clixon_xml_parselval.string = strdup(yytext); return STRING; }
<STRSQ>\' { BEGIN(_YA->ya_lex_state); return *clixon_xml_parsetext; }
%%