commit b8f2056f0393290600d79f69c0ceb0ab3eb29072
parent 13facacc65a3b7895f171618c0e031a437ec53b3
Author: Hiltjo Posthuma <hiltjo@codemadness.org>
Date: Mon, 5 Jun 2017 12:11:28 +0200
wip
Diffstat:
5 files changed, 69 insertions(+), 69 deletions(-)
diff --git a/TODO b/TODO
@@ -1,4 +1,13 @@
+- optimization: for simple patterns use: strstr, strcasestr, strcmp, strcasemp
+
+- combine rules in groups: display: none: display: initial, saves memory:
+ rule1,rule2 { display: none; } vs rule 1{...}rule2{...}
+
+- checkrequest -> allowrequest.
+
- add test-case for global exception rules (no domains), see globalcss init code.
+ also support inverse element hide rule: "~example.com##div.textad".
+- support domain name anchor separately: "||".
- test blocking websocket connections in surf.
- skip protocol part when matching for now, (later add support for websockets).
- for f->matchbegin, matchend etc: make sure to match domain properly, check part of
diff --git a/adblock.c b/adblock.c
@@ -27,17 +27,31 @@ struct filterdomain {
struct filterdomain *next;
};
-struct filterrule {
+struct elementhiderule {
+ /* is exception rule: #@# */
+ int isexception;
+ char *css; /* if non-NULL is CSS rule / hide element rule */
+ struct filterdomain *domains;
+
+ struct elementhiderule *next;
+};
+
+struct blockrule {
/* type: match mask, must be atleast 32-bit, see FilterType enum */
- unsigned long block;
+ unsigned long mask;
+ /* is exception rule: prefix @@ */
+ int isexception;
int matchbegin;
int matchend;
- /* is exception rule: prefix @@ for ABP or #@# for CSS */
- int isexception;
- char *css; /* if non-NULL is CSS rule / hide element rule */
char *uri;
struct filterdomain *domains;
- struct filterrule *next;
+
+ struct blockrule *next;
+};
+
+struct rules {
+ struct blockrule *blockrules;
+ struct elementhiderule *elementhiderules;
};
enum {
@@ -147,27 +161,6 @@ westrdup(const char *s)
return p;
}
-void
-cleanup(void)
-{
- struct filterrule *r;
- struct filterdomain *d;
-
- free(globalcss.data);
-
- for (r = rules; r; r = rules) {
- for (d = r->domains; d; d = r->domains) {
- free(d->domain);
- r->domains = d->next;
- free(d);
- }
- free(r->css);
- free(r->uri);
- rules = r->next;
- free(r);
- }
-}
-
static size_t
string_buffer_realloc(String *s, size_t newsz)
{
@@ -541,36 +534,6 @@ matchrule(struct filterrule *f, const char *uri, const char *type,
return 0;
#endif
-#if 0
- /* DEBUG: test, match if it is a simple pattern */
- char *p;
- p = strchr(f->uri, '*');
- if (!p)
- p = strchr(f->uri, '^');
- if (!p) {
- /* TODO: write a test-case */
- if (f->block & FilterTypeMatchCase) {
- if (f->matchbegin)
- m = strncmp(uri, f->uri, strlen(f->uri)) == 0;
- else if (f->matchend)
- m = strlen(f->uri) <= strlen(uri) &&
- strcmp(&uri[strlen(uri) - strlen(f->uri)], f->uri) == 0;
- else
- m = strstr(uri, f->uri) ? 1 : 0;
- } else {
- if (f->matchbegin)
- m = strncasecmp(uri, f->uri, strlen(f->uri)) == 0;
- else if (f->matchend)
- m = strlen(f->uri) <= strlen(uri) &&
- strcasecmp(&uri[strlen(uri) - strlen(f->uri)], f->uri) == 0;
- else
- m = strcasestr(uri, f->uri) ? 1 : 0;
- }
- /*m = r ? !m : m;*/
- return m;
- }
-#endif
-
r = snprintf(pat, sizeof(pat), "%s%s%s",
f->matchbegin ? "" : "*",
f->uri,
@@ -580,6 +543,11 @@ matchrule(struct filterrule *f, const char *uri, const char *type,
return 0;
}
+ /* DEBUG */
+ if (f->matchbegin) {
+ printf("pat: %s, uri: %s, domain: %s\n", pat, uri, domain);
+ }
+
m = 0;
if (!match(pat, uri, (f->block & FilterTypeMatchCase) ? 0 : 1)) {
#if 0
@@ -629,7 +597,7 @@ parserule(struct filterrule *f, char *s)
}
/* element hiding rule, NOTE: no wildcards are supported,
- "Simplified element hiding syntax" is not supported. */
+ "Simplified element hiding syntax" (legacy) is not supported. */
if ((p = strstr(s, "##"))) {
*p = '\0';
if (parsedomainselement(f, s) < 0)
@@ -734,8 +702,8 @@ loadrules(FILE *fp)
size_t linesiz = 0;
ssize_t n;
int ret;
+ struct rules *rules;
- /* TODO: handle ferror() */
/* load rules */
while ((n = getline(&line, &linesiz, fp)) > 0) {
if (line[n - 1] == '\n')
@@ -755,6 +723,10 @@ loadrules(FILE *fp)
return -1;
}
}
+ if (ferror(fp)) {
+ weprintf("getline: %s\n", strerror(errno));
+ return -1;
+ }
return (rules != NULL);
}
@@ -838,7 +810,7 @@ err:
}
int
-checkrequest(const char *uri, const char *requri)
+allowrequest(const char *uri, const char *requri)
{
char domain[256];
struct filterrule *r;
diff --git a/adblock.h b/adblock.h
@@ -1,4 +1,4 @@
-int checkrequest(const char *, const char *);
+int allowrequest(const char *, const char *);
void cleanup(void);
char *getdocumentcss(const char *);
char *getglobalcss(void);
diff --git a/surf-adblock.c b/surf-adblock.c
@@ -25,11 +25,11 @@ newpage(WebKitWebPage *page)
fprintf(stderr, "surf-adblock: calloc: %s\n", strerror(errno));
return NULL;
}
- p->next = pages;
- pages = p;
-
p->id = webkit_web_page_get_id(page);
p->webpage = page;
+ p->next = pages;
+
+ pages = p;
return p;
}
@@ -60,7 +60,7 @@ documentloaded(WebKitWebPage *wp, Page *p)
}
free(css);
- /* NOTE: globalcss free'd at cleanup() */
+ /* NOTE: globalcss should not be free'd */
}
static gboolean
@@ -68,15 +68,13 @@ sendrequest(WebKitWebPage *wp, WebKitURIRequest *req,
WebKitURIResponse *res, Page *p)
{
const char *uri, *requri;
- gboolean status = FALSE;
if (!webkit_uri_request_get_http_method(req))
- return status;
+ return TRUE; /* TRUE = don't handle any more events */
uri = webkit_web_page_get_uri(p->webpage);
requri = webkit_uri_request_get_uri(req);
- status = checkrequest(uri, requri) ? FALSE : TRUE;
- return status;
+ return allowrequest(uri, requri) ? FALSE : TRUE;
}
static void
diff --git a/tests/tests.c b/tests/tests.c
@@ -1,5 +1,26 @@
#include "../adblock.c"
+void
+cleanup(void)
+{
+ struct filterrule *r;
+ struct filterdomain *d;
+
+ free(globalcss.data);
+
+ for (r = rules; r; r = rules) {
+ for (d = r->domains; d; d = r->domains) {
+ free(d->domain);
+ r->domains = d->next;
+ free(d);
+ }
+ free(r->css);
+ free(r->uri);
+ rules = r->next;
+ free(r);
+ }
+}
+
int
main(void)
{
