summaryrefslogtreecommitdiff
path: root/src/scanners.re
diff options
context:
space:
mode:
Diffstat (limited to 'src/scanners.re')
-rw-r--r--src/scanners.re98
1 files changed, 40 insertions, 58 deletions
diff --git a/src/scanners.re b/src/scanners.re
index 305d1ea..5ec1bf3 100644
--- a/src/scanners.re
+++ b/src/scanners.re
@@ -1,4 +1,4 @@
-#include "bstrlib.h"
+#include <stdlib.h>
/*!re2c
re2c:define:YYCTYPE = "unsigned char";
@@ -36,7 +36,7 @@
htmlcomment = "!--" ([^-\x00]+ | [-][^-\x00]+)* "-->";
- processinginstruction = "?" ([^?>\x00]+ | [?][^>\x00])* "?>";
+ processinginstruction = "?" ([^?>\x00]+ | [?][^>\x00] | [>])* "?>";
declaration = "!" [A-Z]+ spacechar+ [^>\x00]* ">";
@@ -55,11 +55,10 @@
*/
// Try to match URI autolink after first <, returning number of chars matched.
-extern int scan_autolink_uri(bstring s, int pos)
+extern int _scan_autolink_uri(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
scheme [:]([^\x00-\x20<>\\]|escaped_char)*[>] { return (p - start); }
.? { return 0; }
@@ -67,11 +66,10 @@ extern int scan_autolink_uri(bstring s, int pos)
}
// Try to match email autolink after first <, returning num of chars matched.
-extern int scan_autolink_email(bstring s, int pos)
+extern int _scan_autolink_email(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+
[@]
@@ -83,11 +81,10 @@ extern int scan_autolink_email(bstring s, int pos)
}
// Try to match an HTML tag after first <, returning num of chars matched.
-extern int scan_html_tag(bstring s, int pos)
+extern int _scan_html_tag(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
htmltag { return (p - start); }
.? { return 0; }
@@ -96,11 +93,10 @@ extern int scan_html_tag(bstring s, int pos)
// Try to match an HTML block tag including first <,
// returning num of chars matched.
-extern int scan_html_block_tag(bstring s, int pos)
+extern int _scan_html_block_tag(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[<] [/] blocktagname (spacechar | [>]) { return (p - start); }
[<] blocktagname (spacechar | [/>]) { return (p - start); }
@@ -113,11 +109,10 @@ extern int scan_html_block_tag(bstring s, int pos)
// This may optionally be contained in <..>; otherwise
// whitespace and unbalanced right parentheses aren't allowed.
// Newlines aren't ever allowed.
-extern int scan_link_url(bstring s, int pos)
+extern int _scan_link_url(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[ \n]* [<] ([^<>\n\\\x00] | escaped_char | [\\])* [>] { return (p - start); }
[ \n]* (reg_char+ | escaped_char | in_parens_nosp)* { return (p - start); }
@@ -128,11 +123,10 @@ extern int scan_link_url(bstring s, int pos)
// Try to match a link title (in single quotes, in double quotes, or
// in parentheses), returning number of chars matched. Allow one
// level of internal nesting (quotes within quotes).
-extern int scan_link_title(bstring s, int pos)
+extern int _scan_link_title(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
["] (escaped_char|[^"\x00])* ["] { return (p - start); }
['] (escaped_char|[^'\x00])* ['] { return (p - start); }
@@ -142,10 +136,9 @@ extern int scan_link_title(bstring s, int pos)
}
// Match space characters, including newlines.
-extern int scan_spacechars(bstring s, int pos)
+extern int _scan_spacechars(const unsigned char *p)
{
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *start = p; \
/*!re2c
[ \t\n]* { return (p - start); }
. { return 0; }
@@ -153,11 +146,10 @@ extern int scan_spacechars(bstring s, int pos)
}
// Match ATX header start.
-extern int scan_atx_header_start(bstring s, int pos)
+extern int _scan_atx_header_start(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[#]{1,6} ([ ]+|[\n]) { return (p - start); }
.? { return 0; }
@@ -166,10 +158,9 @@ extern int scan_atx_header_start(bstring s, int pos)
// Match sexext header line. Return 1 for level-1 header,
// 2 for level-2, 0 for no match.
-extern int scan_setext_header_line(bstring s, int pos)
+extern int _scan_setext_header_line(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
+ const unsigned char *marker = NULL;
/*!re2c
[=]+ [ ]* [\n] { return 1; }
[-]+ [ ]* [\n] { return 2; }
@@ -180,11 +171,10 @@ extern int scan_setext_header_line(bstring s, int pos)
// Scan a horizontal rule line: "...three or more hyphens, asterisks,
// or underscores on a line by themselves. If you wish, you may use
// spaces between the hyphens or asterisks."
-extern int scan_hrule(bstring s, int pos)
+extern int _scan_hrule(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
([*][ ]*){3,} [ \t]* [\n] { return (p - start); }
([_][ ]*){3,} [ \t]* [\n] { return (p - start); }
@@ -194,11 +184,10 @@ extern int scan_hrule(bstring s, int pos)
}
// Scan an opening code fence.
-extern int scan_open_code_fence(bstring s, int pos)
+extern int _scan_open_code_fence(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[`]{3,} / [^`\n\x00]*[\n] { return (p - start); }
[~]{3,} / [^~\n\x00]*[\n] { return (p - start); }
@@ -207,29 +196,22 @@ extern int scan_open_code_fence(bstring s, int pos)
}
// Scan a closing code fence with length at least len.
-extern int scan_close_code_fence(bstring s, int pos, int len)
+extern int _scan_close_code_fence(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
- ([`]{3,} | [~]{3,}) / spacechar* [\n]
- { if (p - start > len) {
- return (p - start);
- } else {
- return 0;
- } }
+ ([`]{3,} | [~]{3,}) / spacechar* [\n] { return (p - start); }
.? { return 0; }
*/
}
// Scans an entity.
// Returns number of chars matched.
-extern int scan_entity(bstring s, int pos)
+extern int _scan_entity(const unsigned char *p)
{
- unsigned char * marker = NULL;
- unsigned char * p = &(s->data[pos]);
- unsigned char * start = p;
+ const unsigned char *marker = NULL;
+ const unsigned char *start = p;
/*!re2c
[&] ([#] ([Xx][A-Fa-f0-9]{1,8}|[0-9]{1,8}) |[A-Za-z][A-Za-z0-9]{1,31} ) [;]
{ return (p - start); }
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-19 14:54:43 +0000