Merge pull request #126 from iresty/optimize-wildcard

optimize: optimized pattern `.*`, which can be used prefix matching.

README.md

@@ -125,7 +125,7 @@ R3Node * matched_node = r3_tree_match_entry(n, entry);
 
 **Release Memory**
 
-To release the memory, you may call `r3_tree_free(R3Node *tree)` to release the whole tree structure, 
+To release the memory, you may call `r3_tree_free(R3Node *tree)` to release the whole tree structure,
 `node*`, `edge*`, `route*` objects that were inserted into the tree will be freed.
 
 
@@ -189,7 +189,7 @@ translator, which translates simple patterns into small & fast scanners.
 
 By using this method, r3 reduces the matching overhead of pcre library.
 
-Optimized patterns are: `[a-z]+`, `[0-9]+`, `\d+`, `\w+`, `[^/]+` or `[^-]+`
+Optimized patterns are: `[a-z]+`, `[0-9]+`, `\d+`, `\w+`, `[^/]+`, `[^-]+` or `.*`.
 
 Slugs without specified regular expression will be compiled into the `[^/]+` pattern. therefore, it's optimized too.
 
@@ -337,10 +337,10 @@ Use case in PHP
 ```php
 // Here is the paths data structure
 $paths = [
-    '/blog/post/{id}' => [ 'controller' => 'PostController' , 'action' => 'item'   , 'method'   => 'GET' ] , 
-    '/blog/post'      => [ 'controller' => 'PostController' , 'action' => 'list'   , 'method'   => 'GET' ] , 
-    '/blog/post'      => [ 'controller' => 'PostController' , 'action' => 'create' , 'method' => 'POST' ]  , 
-    '/blog'           => [ 'controller' => 'BlogController' , 'action' => 'list'   , 'method'   => 'GET' ] , 
+    '/blog/post/{id}' => [ 'controller' => 'PostController' , 'action' => 'item'   , 'method'   => 'GET' ] ,
+    '/blog/post'      => [ 'controller' => 'PostController' , 'action' => 'list'   , 'method'   => 'GET' ] ,
+    '/blog/post'      => [ 'controller' => 'PostController' , 'action' => 'create' , 'method' => 'POST' ]  ,
+    '/blog'           => [ 'controller' => 'BlogController' , 'action' => 'list'   , 'method'   => 'GET' ] ,
 ];
 $rs = r3_compile($paths, 'persisten-table-id');
 $ret = r3_dispatch($rs, '/blog/post/3' );

include/r3.h

@@ -208,7 +208,8 @@ int r3_pattern_to_opcode(const char * pattern, unsigned int len);
 
 enum { NODE_COMPARE_STR, NODE_COMPARE_PCRE, NODE_COMPARE_OPCODE };
 
-enum { OP_EXPECT_MORE_DIGITS = 1, OP_EXPECT_MORE_WORDS, OP_EXPECT_NOSLASH, OP_EXPECT_NODASH, OP_EXPECT_MORE_ALPHA };
+enum { OP_EXPECT_MORE_DIGITS = 1, OP_EXPECT_MORE_WORDS, OP_EXPECT_NOSLASH,
+       OP_EXPECT_NODASH, OP_EXPECT_MORE_ALPHA, OP_GREEDY_ANY};
 
 
 

src/node.c

@@ -223,7 +223,7 @@ int r3_tree_compile_patterns(R3Node * n, char **errstr) {
     free(n->combined_pattern);
     n->combined_pattern = cpat;
 
-    const char *pcre_error;
+    const char *pcre_error = NULL;
     int pcre_erroffset;
     unsigned int option_bits = 0;
 
@@ -250,7 +250,7 @@ int r3_tree_compile_patterns(R3Node * n, char **errstr) {
         pcre_free_study(n->pcre_extra);
     }
     n->pcre_extra = pcre_study(n->pcre_pattern, 0, &pcre_error);
-    if (!n->pcre_extra) {
+    if (!n->pcre_extra && pcre_error) {
         if (errstr) {
             int r = asprintf(errstr, "PCRE study failed at offset %s, pattern: %s", pcre_error, n->combined_pattern);
             if (r) {};
@@ -312,18 +312,32 @@ R3Node * r3_tree_matchl(const R3Node * n, const char * path, unsigned int path_l
                 case OP_EXPECT_NODASH:
                     while (*pp != '-' && pp < pp_end) pp++;
                     break;
+                case OP_GREEDY_ANY:
+                    while (*pp != '\n' && pp < pp_end) pp++;
+                    break;
             }
+
             // check match
-            if ((pp - path) > 0) {
+            if (e->opcode != OP_GREEDY_ANY) {
+                if ((pp - path) > 0) {
+                    if (entry) {
+                        str_array_append(&entry->vars , path, pp - path);
+                    }
+                    restlen = pp_end - pp;
+                    if (!restlen) {
+                        return e->child && e->child->endpoint ? e->child : NULL;
+                    }
+                    return r3_tree_matchl(e->child, pp, restlen, entry);
+                }
+
+            } else {
                 if (entry) {
                     str_array_append(&entry->vars , path, pp - path);
                 }
-                restlen = pp_end - pp;
-                if (!restlen) {
-                    return e->child && e->child->endpoint ? e->child : NULL;
-                }
-                return r3_tree_matchl(e->child, pp, restlen, entry);
+
+                return e->child && e->child->endpoint ? e->child : NULL;
             }
+
             e++;
         }
     }

src/str.c

@@ -43,6 +43,9 @@ int r3_pattern_to_opcode(const char * pattern, unsigned int len) {
     if ( strncmp(pattern, "[^-]+", len) == 0 ) {
         return OP_EXPECT_NODASH;
     }
+    if ( strncmp(pattern, ".*", len) == 0 ) {
+        return OP_GREEDY_ANY;
+    }
     return 0;
 }
 

tests/CMakeLists.txt

@@ -21,6 +21,7 @@ add_r3_test(check_str_array check_str_array.c)
 add_r3_test(check_host check_host.c)
 add_r3_test(check_http_scheme check_http_scheme.c)
 add_r3_test(check_remote_addr check_remote_addr.c)
+add_r3_test(check_routes2 check_routes2.c)
 
 
 add_executable(bench bench.c)

tests/Makefile.am

@@ -36,6 +36,9 @@ check_remote_addr_SOURCES = check_remote_addr.c
 TESTS += check_http_scheme
 check_http_scheme_SOURCES = check_http_scheme.c
 
+TESTS += check_routes2
+check_routes2_SOURCES = check_routes2.c
+
 
 if ENABLE_JSON
 TESTS += check_json

tests/check_routes2.c

@@ -0,0 +1,65 @@
+#include "config.h"
+#include <stdio.h>
+#include <check.h>
+#include <stdlib.h>
+#include <assert.h>
+#include "r3.h"
+#include "r3_slug.h"
+
+START_TEST (greedy_pattern)
+{
+    R3Node * n      = r3_tree_create(10);
+    match_entry * entry;
+    R3Route *matched_route;
+
+    char * uri0 = "/foo{:.*}";
+    r3_tree_insert_routel(n, 0, uri0, strlen(uri0), &uri0);
+
+    char * err = NULL;
+    r3_tree_compile(n, &err);
+    ck_assert(err == NULL);
+
+    entry = match_entry_create("/foo/bar");
+    matched_route = r3_tree_match_route(n, entry);
+    ck_assert(matched_route != NULL);
+    ck_assert(matched_route->data == &uri0);
+
+    // fixme: should match
+
+    // entry = match_entry_create("/foo");
+    // matched_route = r3_tree_match_route(n, entry);
+    // ck_assert(matched_route != NULL);
+    // ck_assert(matched_route->data == &uri0);
+
+    entry = match_entry_create("/foo/");
+    matched_route = r3_tree_match_route(n, entry);
+    ck_assert(matched_route != NULL);
+    ck_assert(matched_route->data == &uri0);
+
+    entry = match_entry_create("/foo/bar/foo/mmasdfasdfasd/f/asdf/as/df");
+    matched_route = r3_tree_match_route(n, entry);
+    ck_assert(matched_route != NULL);
+    ck_assert(matched_route->data == &uri0);
+
+    r3_tree_free(n);
+}
+END_TEST
+
+Suite* r3_suite (void) {
+    Suite *suite = suite_create("r3 routes2 tests");
+    TCase *tcase = tcase_create("testcase");
+    tcase_add_test(tcase, greedy_pattern);
+    suite_add_tcase(suite, tcase);
+    return suite;
+}
+
+int main (int argc, char *argv[]) {
+    int number_failed;
+    Suite *suite = r3_suite();
+    SRunner *runner = srunner_create(suite);
+    srunner_run_all(runner, CK_NORMAL);
+    number_failed = srunner_ntests_failed(runner);
+    srunner_free(runner);
+    return number_failed;
+}
+