Merge pull request #288 from tiehuis/master

Add RegexSet functionality to C API

Author: BurntSushi

regex-capi/README.md

@@ -99,7 +99,5 @@ There are a few things missing from the C API that are present in the Rust API.
 There's no particular (known) reason why they don't, they just haven't been
 implemented yet.
 
-* RegexSet, which permits matching multiple regular expressions simultaneously
-  in a single linear time search.
 * Splitting a string by a regex.
 * Replacing regex matches in a string with some other text.

regex-capi/ctest/test.c

@@ -331,6 +331,191 @@ bool test_compile_error_size_limit() {
     return passed;
 }
 
+bool test_regex_set_matches() {
+
+#define PAT_COUNT 6
+
+    bool passed = true;
+    const char *patterns[] = {
+        "foo", "barfoo", "\\w+", "\\d+", "foobar", "bar"
+    };
+    const size_t patterns_lengths[] = {
+        3, 6, 3, 3, 6, 3
+    };
+
+    rure_error *err = rure_error_new();
+    rure_set *re = rure_compile_set((const uint8_t **) patterns,
+                                    patterns_lengths,
+                                    PAT_COUNT,
+                                    0,
+                                    NULL,
+                                    err);
+    if (re == NULL) {
+        passed = false;
+        goto done2;
+    }
+
+    if (rure_set_len(re) != PAT_COUNT) {
+        passed = false;
+        goto done1;
+    }
+
+    if (!rure_set_is_match(re, (const uint8_t *) "foobar", 6, 0)) {
+        passed = false;
+        goto done1;
+    }
+
+    if (rure_set_is_match(re, (const uint8_t *) "", 0, 0)) {
+        passed = false;
+        goto done1;
+    }
+
+    bool matches[PAT_COUNT];
+    if (!rure_set_matches(re, (const uint8_t *) "foobar", 6, 0, matches)) {
+        passed = false;
+        goto done1;
+    }
+
+    const bool match_target[] = {
+        true, false, true, false, true, true
+    };
+
+    int i;
+    for (i = 0; i < PAT_COUNT; ++i) {
+        if (matches[i] != match_target[i]) {
+            passed = false;
+            goto done1;
+        }
+    }
+
+done1:
+    rure_set_free(re);
+done2:
+    rure_error_free(err);
+    return passed;
+
+#undef PAT_COUNT
+}
+
+bool test_regex_set_match_start() {
+
+#define PAT_COUNT 3
+
+    bool passed = true;
+    const char *patterns[] = {
+        "foo", "bar", "fooo"
+    };
+    const size_t patterns_lengths[] = {
+        3, 3, 4
+    };
+
+    rure_error *err = rure_error_new();
+    rure_set *re = rure_compile_set((const uint8_t **) patterns,
+                                    patterns_lengths,
+                                    PAT_COUNT,
+                                    0,
+                                    NULL,
+                                    err);
+    if (re == NULL) {
+        passed = false;
+        goto done2;
+    }
+
+    if (rure_set_len(re) != PAT_COUNT) {
+        passed = false;
+        goto done1;
+    }
+
+    if (rure_set_is_match(re, (const uint8_t *) "foobiasdr", 7, 2)) {
+        passed = false;
+        goto done1;
+    }
+
+    {
+        bool matches[PAT_COUNT];
+        if (!rure_set_matches(re, (const uint8_t *) "fooobar", 8, 0, matches)) {
+            passed = false;
+            goto done1;
+        }
+
+        const bool match_target[] = {
+            true, true, true
+        };
+
+        int i;
+        for (i = 0; i < PAT_COUNT; ++i) {
+            if (matches[i] != match_target[i]) {
+                passed = false;
+                goto done1;
+            }
+        }
+    }
+
+    {
+        bool matches[PAT_COUNT];
+        if (!rure_set_matches(re, (const uint8_t *) "fooobar", 7, 1, matches)) {
+            passed = false;
+            goto done1;
+        }
+
+        const bool match_target[] = {
+            false, true, false
+        };
+
+        int i;
+        for (i = 0; i < PAT_COUNT; ++i) {
+            if (matches[i] != match_target[i]) {
+                passed = false;
+                goto done1;
+            }
+        }
+    }
+
+done1:
+    rure_set_free(re);
+done2:
+    rure_error_free(err);
+    return passed;
+
+#undef PAT_COUNT
+}
+
+bool test_regex_set_options() {
+
+    bool passed = true;
+    rure_options *opts = rure_options_new();
+    rure_options_size_limit(opts, 0);
+    rure_error *err = rure_error_new();
+
+    const char *patterns[] = { "\\w{100}" };
+    const size_t patterns_lengths[] = { 8 };
+
+    rure_set *re = rure_compile_set((const uint8_t **) patterns, patterns_lengths,
+                                    1, 0, opts, err);
+    if (re != NULL) {
+        if (DEBUG) {
+            fprintf(stderr,
+                    "[test_compile_error_size_limit] "
+                    "expected NULL regex pointer, but got non-NULL pointer\n");
+        }
+        passed = false;
+        rure_set_free(re);
+    }
+    const char *msg = rure_error_message(err);
+    if (NULL == strstr(msg, "exceeds size")) {
+        if (DEBUG) {
+            fprintf(stderr,
+                    "[test_compile_error] "
+                    "expected an 'exceeds size' error message, but "
+                    "got this instead: '%s'\n", msg);
+        }
+        passed = false;
+    }
+    rure_options_free(opts);
+    rure_error_free(err);
+    return passed;
+}
+
 void run_test(bool (test)(), const char *name, bool *passed) {
     if (!test()) {
         *passed = false;
@@ -353,6 +538,9 @@ int main() {
     run_test(test_compile_error, "test_compile_error", &passed);
     run_test(test_compile_error_size_limit, "test_compile_error_size_limit",
              &passed);
+    run_test(test_regex_set_matches, "test_regex_set_match", &passed);
+    run_test(test_regex_set_options, "test_regex_set_options", &passed);
+    run_test(test_regex_set_match_start, "test_regex_set_match_start", &passed);
 
     if (!passed) {
         exit(1);

regex-capi/include/rure.h

@@ -28,6 +28,13 @@ extern "C" {
  */
 typedef struct rure rure;
 
+/*
+ * rure_set is the type of a set of compiled regular expressions.
+ *
+ * A rure can be safely used from multiple threads simultaneously.
+ */
+typedef struct rure_set rure_set;
+
 /*
  * rure_options is the set of non-flag configuration options for compiling
  * a regular expression. Currently, only two options are available: setting
@@ -165,7 +172,7 @@ rure *rure_compile(const uint8_t *pattern, size_t length,
 /*
  * rure_free frees the given compiled regular expression.
  *
- * This must be called at most once.
+ * This must be called at most once for any rure.
  */
 void rure_free(rure *re);
 
@@ -446,6 +453,90 @@ void rure_options_size_limit(rure_options *options, size_t limit);
  */
 void rure_options_dfa_size_limit(rure_options *options, size_t limit);
 
+/*
+ * rure_compile_set compiles the given list of patterns into a single regular
+ * expression which can be matched in a linear-scan. Each pattern in patterns
+ * must be valid UTF-8 and the length of each pattern in patterns corresponds
+ * to a byte length in patterns_lengths.
+ *
+ * The number of patterns to compile is specified by patterns_count. patterns
+ * must contain at least this many entries.
+ *
+ * flags is a bitfield. Valid values are constants declared with prefix
+ * RURE_FLAG_.
+ *
+ * options contains non-flag configuration settings. If it's NULL, default
+ * settings are used. options may be freed immediately after a call to
+ * rure_compile.
+ *
+ * error is set if there was a problem compiling the pattern.
+ *
+ * The compiled expression set returned may be used from multiple threads.
+ */
+rure_set *rure_compile_set(const uint8_t **patterns,
+                           const size_t *patterns_lengths,
+                           size_t patterns_count,
+                           uint32_t flags,
+                           rure_options *options,
+                           rure_error *error);
+
+/*
+ * rure_set_free frees the given compiled regular expression set.
+ *
+ * This must be called at most once for any rure_set.
+ */
+void rure_set_free(rure_set *re);
+
+/*
+ * rure_is_match returns true if and only if any regexes within the set
+ * match anywhere in the haystack. Once a match has been located, the
+ * matching engine will quit immediately.
+ *
+ * haystack may contain arbitrary bytes, but ASCII compatible text is more
+ * useful. UTF-8 is even more useful. Other text encodings aren't supported.
+ * length should be the number of bytes in haystack.
+ *
+ * start is the position at which to start searching. Note that setting the
+ * start position is distinct from incrementing the pointer, since the regex
+ * engine may look at bytes before the start position to determine match
+ * information. For example, if the start position is greater than 0, then the
+ * \A ("begin text") anchor can never match.
+ */
+bool rure_set_is_match(rure_set *re, const uint8_t *haystack, size_t length,
+                       size_t start);
+
+/*
+ * rure_set_matches compares each regex in the set against the haystack and
+ * modifies matches with the match result of each pattern. Match results are
+ * ordered in the same way as the rure_set was compiled. For example,
+ * index 0 of matches corresponds to the first pattern passed to
+ * `rure_compile_set`.
+ *
+ * haystack may contain arbitrary bytes, but ASCII compatible text is more
+ * useful. UTF-8 is even more useful. Other text encodings aren't supported.
+ * length should be the number of bytes in haystack.
+ *
+ * start is the position at which to start searching. Note that setting the
+ * start position is distinct from incrementing the pointer, since the regex
+ * engine may look at bytes before the start position to determine match
+ * information. For example, if the start position is greater than 0, then the
+ * \A ("begin text") anchor can never match.
+ *
+ * matches must be greater than or equal to the number of patterns the
+ * rure_set was compiled with.
+ *
+ * Only use this function if you specifically need to know which regexes
+ * matched within the set. To determine if any of the regexes matched without
+ * caring which, use rure_set_is_match.
+ */
+bool rure_set_matches(rure_set *re, const uint8_t *haystack, size_t length,
+                      size_t start, bool *matches);
+
+/*
+ * rure_set_len returns the number of patterns rure_set was compiled with.
+ */
+size_t rure_set_len(rure_set *re);
+
 /*
  * rure_error_new allocates space for an error.
  *