mirrors/netsurf
1
0
Fork 0
mirror of https://github.com/netsurf-browser/netsurf synced 2025-02-07 01:55:00 +03:00
Code Issues Packages Projects Releases Wiki Activity

Contribution from Philip Boulain <prb@ecs.soton.ac.uk>:

Browse Source 
This makes url_normalize take care of whitespace in a fairly useful way,
consistent with other browsers:

 - Leading and trailing whitespace is trimmed
 - Internal whitespace is urlescaped

For example,
 "  http://www.google.co.uk/search?q=hello world  "
becomes
 "http://www.google.co.uk/search?q=hello%20world"

Explicit trailing whitespace, e.g. "...hello world%20", is left alone.

The upshot is that if you sloppily copy-paste a URL from IRC or whatnot
into the address bar, NetSurf no longer silently ignores you if you
caught some adjacent whitespace.

svn path=/trunk/netsurf/; revision=4198
This commit is contained in:
John Tytgat 2008-05-25 15:04:22 +00:00
parent 497372c257
commit 3f6d2a9f0c
2 changed files with 85 additions and 44 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

117
utils/url.c
View File

@ -127,41 +127,69 @@ url_func_result url_normalize(const char *url, char **result)
{ {
char c; char c;
int m; int m;
int i; size_t i;
size_t len; size_t len;
size_t bufsize;
char* norm;
bool http = false; bool http = false;
regmatch_t match[10]; regmatch_t match[10];
*result = NULL; *result = NULL;
if ((m = regexec(&url_re, url, 10, match, 0))) { /* skip past any leading whitespace (likely if URL was copy-pasted) */
while (isspace(*url))
url++;
/* allocate sufficiently large buffer for new URL */
len = strlen(url);
bufsize = len + sizeof("http://")-1 + sizeof("/")-1 + 1; /* 'http://' + '/' + '\0' */
/* work out how much extra to leave for internal whitespace */
for(i = 0; i < len; i++) {
if(isspace(url[i])) bufsize += 2; /* ' ' -> '%20' */
}
if ((norm = malloc(bufsize)) == NULL) {
LOG(("malloc failed"));
return URL_FUNC_NOMEM;
}
*result = norm;
strcpy(norm, url);
/* truncate trailing whitespace (significant should be uriencoded) */
for (i = len - 1; (i > 0) && isspace(norm[i]); i--) {
norm[i] = '\0';
len--;
}
/* encode any remaining (internal) whitespace */
for (i = 0; i < len; i++) {
if(isspace(norm[i])) {
char space = norm[i];
memmove(norm + i + 2, norm + i, 1 + len - i);
len += 2;
norm[ i] = '%';
norm[++i] = digit2lowcase_hex(space >> 4);
norm[++i] = digit2lowcase_hex(space & 0xf);
}
}
/* finally verify that it's actually an URL we're working on
* (RFC regex too fussy to tolerate above WSP problems) */
if ((m = regexec(&url_re, norm, 10, match, 0))) {
LOG(("url '%s' failed to match regex", url)); LOG(("url '%s' failed to match regex", url));
return URL_FUNC_FAILED; return URL_FUNC_FAILED;
} }
len = strlen(url);
if (match[URL_RE_SCHEME].rm_so == -1) { if (match[URL_RE_SCHEME].rm_so == -1) {
/* scheme missing: add http:// and reparse */ /* scheme missing: add http:// and reparse */
/* LOG(("scheme missing: using http"));*/ /* LOG(("scheme missing: using http"));*/
if ((*result = malloc(len + 13)) == NULL) { memmove(norm + sizeof("http://")-1, norm, len + 1);
LOG(("malloc failed")); memcpy(norm, "http://", sizeof("http://")-1); /* do NOT copy null */
return URL_FUNC_NOMEM; len += 7;
} if ((m = regexec(&url_re, norm, 10, match, 0))) {
strcpy(*result, "http://"); LOG(("url '%s' failed to match regex", norm));
strcpy(*result + sizeof("http://")-1, url); free(norm);
if ((m = regexec(&url_re, *result, 10, match, 0))) {
LOG(("url '%s' failed to match regex", (*result)));
free(*result);
return URL_FUNC_FAILED; return URL_FUNC_FAILED;
} }
len += sizeof("http://")-1;
} else {
if ((*result = malloc(len + 6)) == NULL) {
LOG(("malloc failed"));
return URL_FUNC_NOMEM;
}
strcpy(*result, url);
} }
/*for (unsigned int i = 0; i != 10; i++) { /*for (unsigned int i = 0; i != 10; i++) {
@ -177,22 +205,22 @@ url_func_result url_normalize(const char *url, char **result)
if (match[URL_RE_SCHEME].rm_so != -1) { if (match[URL_RE_SCHEME].rm_so != -1) {
for (i = match[URL_RE_SCHEME].rm_so; for (i = match[URL_RE_SCHEME].rm_so;
i != match[URL_RE_SCHEME].rm_eo; i++) i != match[URL_RE_SCHEME].rm_eo; i++)
(*result)[i] = tolower((*result)[i]); norm[i] = tolower(norm[i]);
if (match[URL_RE_SCHEME].rm_eo == 4 if (match[URL_RE_SCHEME].rm_eo == 4
&& (*result)[0] == 'h' && norm[0] == 'h'
&& (*result)[1] == 't' && norm[1] == 't'
&& (*result)[2] == 't' && norm[2] == 't'
&& (*result)[3] == 'p') && norm[3] == 'p')
http = true; http = true;
} }
/* make empty path into "/" */ /* make empty path into "/" */
if (match[URL_RE_PATH].rm_so != -1 && if (match[URL_RE_PATH].rm_so != -1 &&
match[URL_RE_PATH].rm_so == match[URL_RE_PATH].rm_eo) { match[URL_RE_PATH].rm_so == match[URL_RE_PATH].rm_eo) {
memmove((*result) + match[URL_RE_PATH].rm_so + 1, memmove(norm + match[URL_RE_PATH].rm_so + 1,
(*result) + match[URL_RE_PATH].rm_so, norm + match[URL_RE_PATH].rm_so,
len - match[URL_RE_PATH].rm_so + 1); len - match[URL_RE_PATH].rm_so + 1);
(*result)[match[URL_RE_PATH].rm_so] = '/'; norm[match[URL_RE_PATH].rm_so] = '/';
len++; len++;
} }
@ -200,45 +228,45 @@ url_func_result url_normalize(const char *url, char **result)
if (match[URL_RE_AUTHORITY].rm_so != -1) { if (match[URL_RE_AUTHORITY].rm_so != -1) {
for (i = match[URL_RE_AUTHORITY].rm_so; for (i = match[URL_RE_AUTHORITY].rm_so;
i != match[URL_RE_AUTHORITY].rm_eo; i++) { i != match[URL_RE_AUTHORITY].rm_eo; i++) {
if ((*result)[i] == ':') { if (norm[i] == ':' && (i + 3) < len) {
if (http && (*result)[i + 1] == '8' && if (http && norm[i + 1] == '8' &&
(*result)[i + 2] == '0' && norm[i + 2] == '0' &&
i + 3 == i + 3 ==
match[URL_RE_AUTHORITY].rm_eo) { match[URL_RE_AUTHORITY].rm_eo) {
memmove((*result) + i, memmove(norm + i,
(*result) + i + 3, norm + i + 3,
len - len -
match[URL_RE_AUTHORITY]. match[URL_RE_AUTHORITY].
rm_eo); rm_eo);
len -= 3; len -= 3;
(*result)[len] = '\0'; norm[len] = '\0';
} else if (i + 1 == match[4].rm_eo) { } else if (i + 1 == match[4].rm_eo) {
memmove((*result) + i, memmove(norm + i,
(*result) + i + 1, norm + i + 1,
len - len -
match[URL_RE_AUTHORITY]. match[URL_RE_AUTHORITY].
rm_eo); rm_eo);
len--; len--;
(*result)[len] = '\0'; norm[len] = '\0';
} }
break; break;
} }
(*result)[i] = tolower((*result)[i]); norm[i] = tolower(norm[i]);
} }
} }
/* unescape non-"reserved" escaped characters */ /* unescape non-"reserved" escaped characters */
for (i = 0; (unsigned)i != len; i++) { for (i = 0; i + 2 < len; i++) {
if ((*result)[i] != '%') if (norm[i] != '%')
continue; continue;
c = tolower((*result)[i + 1]); c = tolower(norm[i + 1]);
if ('0' <= c && c <= '9') if ('0' <= c && c <= '9')
m = 16 * (c - '0'); m = 16 * (c - '0');
else if ('a' <= c && c <= 'f') else if ('a' <= c && c <= 'f')
m = 16 * (c - 'a' + 10); m = 16 * (c - 'a' + 10);
else else
continue; continue;
c = tolower((*result)[i + 2]); c = tolower(norm[i + 2]);
if ('0' <= c && c <= '9') if ('0' <= c && c <= '9')
m += c - '0'; m += c - '0';
else if ('a' <= c && c <= 'f') else if ('a' <= c && c <= 'f')
@ -252,11 +280,12 @@ url_func_result url_normalize(const char *url, char **result)
continue; continue;
} }
(*result)[i] = m; norm[i] = m;
memmove((*result) + i + 1, (*result) + i + 3, len - i - 2); memmove(norm + i + 1, norm + i + 3, len - i - 2);
len -= 2; len -= 2;
} }
/* norm and *result point to same memory, so just return ok */
return URL_FUNC_OK; return URL_FUNC_OK;
} }

12
utils/utils.h
View File

@ -25,6 +25,7 @@
#include <stdlib.h> #include <stdlib.h>
#include <sys/types.h> #include <sys/types.h>
#include <regex.h> #include <regex.h>
#include <assert.h>
#ifndef NOF_ELEMENTS #ifndef NOF_ELEMENTS
#define NOF_ELEMENTS(array) (sizeof(array)/sizeof(*(array))) #define NOF_ELEMENTS(array) (sizeof(array)/sizeof(*(array)))
@ -71,6 +72,17 @@ char *strcasestr(const char *haystack, const char *needle);
#endif #endif
unsigned int wallclock(void); unsigned int wallclock(void);
/**
* Return a hex digit for the given numerical value.
*
* \return character in range 0-9a-f
*/
inline static char digit2lowcase_hex(unsigned char digit) {
assert(digit < 16);
return "0123456789abcdef"[digit];
}
/* Platform specific functions */ /* Platform specific functions */
void die(const char * const error); void die(const char * const error);
void warn_user(const char *warning, const char *detail); void warn_user(const char *warning, const char *detail);