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

Removing newlines from XML data [https://github.com/clicon/clixon/issues/65]

Browse source
This commit is contained in:
Olof hagsand 2019-01-16 23:17:07 +01:00
parent 26a75a4c50
commit 03ac48ab2b
6 changed files with 44 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

5
lib/src/clixon_xml.c
View file

@ -1483,7 +1483,7 @@ _xml_parse(const char *str,
{
int retval = -1;
struct xml_parse_yacc_arg ya = {0,};
cxobj *x;
if (strlen(str) == 0)
return 0; /* OK */
if (xt == NULL){
@ -1501,6 +1501,9 @@ _xml_parse(const char *str,
goto done;
if (clixon_xml_parseparse(&ya) != 0) /* yacc returns 1 on error */
goto done;
x = NULL;
while ((x = xml_find_type(xt, NULL, "body", CX_BODY)) != NULL)
xml_purge(x);
/* Verify namespaces after parsing */
if (xml_apply0(xt, CX_ELMNT, xml_localname_check, NULL) < 0)
goto done;

8
lib/src/clixon_xml_parse.l
View file

@ -105,8 +105,8 @@ ncname {namestart}{namechar}*
%%
<START,TEXTDECL>[ \t] ;
<START,STATEA,CMNT,TEXTDECL>\n { _YA->ya_linenum++; }
<START,STATEA,CMNT,TEXTDECL>\r
<START,CMNT,TEXTDECL>\n { _YA->ya_linenum++; }
<START,CMNT,TEXTDECL>\r
<START>{ncname} { clixon_xml_parselval.string = strdup(yytext);
return NAME; /* rather be catch-all */
@ -133,6 +133,10 @@ ncname {namestart}{namechar}*
<STATEA>"<?" { BEGIN(PIDECL); return BQMARK; }
<STATEA>\< { BEGIN(START); return *clixon_xml_parsetext; }
<STATEA>& { _YA->ya_lex_state =STATEA;BEGIN(AMPERSAND);}
<STATEA>[ \t] { clixon_xml_parselval.string = yytext;return WHITESPACE; }
<STATEA>\r\n { clixon_xml_parselval.string = "\n";return WHITESPACE; }
<STATEA>\r { clixon_xml_parselval.string = "\n";return WHITESPACE; }
<STATEA>\n { clixon_xml_parselval.string = yytext; _YA->ya_linenum++;return WHITESPACE; }
<STATEA>. { clixon_xml_parselval.string = yytext; return CHARDATA; }
/* @see xml_chardata_encode */

11
lib/src/clixon_xml_parse.y
View file

@ -41,7 +41,7 @@
%start document
%token <string> NAME CHARDATA STRING
%token <string> NAME CHARDATA WHITESPACE STRING
%token MY_EOF
%token VER ENC SD
%token BSLASH ESLASH
@ -357,12 +357,13 @@ prolog : xmldcl misclist
;
misclist : misclist misc { clicon_debug(2, "misclist->misclist misc"); }
| { clicon_debug(2, "misclist->"); }
| { clicon_debug(2, "misclist->"); }
;
/* [27] Misc ::= Comment | PI | S */
misc : comment { clicon_debug(2, "misc->comment"); }
| pi { clicon_debug(2, "misc->pi"); }
misc : comment { clicon_debug(2, "misc->comment"); }
| pi { clicon_debug(2, "misc->pi"); }
| WHITESPACE { clicon_debug(2, "misc->white space"); }
;
xmldcl : BXMLDCL verinfo encodingdecl sddecl EQMARK
@ -424,6 +425,8 @@ content : element { clicon_debug(2, "content -> element"); }
| pi { clicon_debug(2, "content -> pi"); }
| CHARDATA { if (xml_parse_content(_YA, $1) < 0) YYABORT;
clicon_debug(2, "content -> CHARDATA %s", $1); }
| WHITESPACE { if (xml_parse_content(_YA, $1) < 0) YYABORT;
clicon_debug(2, "content -> WHITESPACE %s", $1); }
| { clicon_debug(2, "content -> "); }
;