Add support for regress rust crate

This adds support for the new regress regex engine.

Author: ridiculousfish

README.md

@@ -31,6 +31,7 @@ The following regex engines are supported and covered by the tool:
 - [Tre](https://github.com/laurikari/tre)
 - [PCRE2](http://www.pcre.org)
 - [Rust regex crate](https://doc.rust-lang.org/regex/regex/index.html)
+- [Regress regex crate](https://docs.rs/regress/)
 
 The engines are built from their sources. In the case an installed engine should be used,
 the corresponding cmake variable `INCLUDE_<name>` has to be set to `system`. The configuration script

src/main.c

@@ -33,6 +33,7 @@ static struct engines engines [] = {
     {.name = "hscan",       .find_all = hs_find_all},
 #endif
     {.name = "rust_regex",  .find_all = rust_find_all},
+    {.name = "rust_regrs",  .find_all = regress_find_all},
 };
 
 static char * regex [] = {

src/main.h

@@ -40,6 +40,7 @@ int onig_find_all(char* pattern, char* subject, int subject_len, int repeat, str
 int hs_find_all(char * pattern, char * subject, int subject_len, int repeat, struct result * res);
 #endif
 int rust_find_all(char * pattern, char * subject, int subject_len, int repeat, struct result * res);
+int regress_find_all(char * pattern, char * subject, int subject_len, int repeat, struct result * res);
 
 #ifdef __cplusplus
 }

src/rust.c

@@ -35,3 +35,37 @@ int rust_find_all(char * pattern, char * subject, int subject_len, int repeat, s
 
     return 0;
 }
+
+int regress_find_all(char *pattern, char *subject, int subject_len, int repeat, struct result *res)
+{
+    TIME_TYPE start, end;
+    int found = 0;
+
+    struct Regress const *regex_hdl = regress_new(pattern);
+    if (regex_hdl == NULL)
+    {
+        fprintf(stderr, "ERROR: Unable to compile pattern \"%s\"\n", pattern);
+        return -1;
+    }
+
+    double *times = calloc(repeat, sizeof(double));
+    int const times_len = repeat;
+
+    do
+    {
+        GET_TIME(start);
+        found = regress_matches(regex_hdl, (uint8_t *)subject, subject_len);
+        GET_TIME(end);
+
+        times[repeat - 1] = TIME_DIFF_IN_MS(start, end);
+
+    } while (--repeat > 0);
+
+    res->matches = found;
+    get_mean_and_derivation(times, times_len, res);
+
+    regress_free(regex_hdl);
+    free(times);
+
+    return 0;
+}

src/rust/Cargo.toml

@@ -9,6 +9,7 @@ crate-type = ["staticlib"]
 
 [dependencies]
 regex = "1.3.7"
+regress = "0.1"
 libc = "0.2.19"
 
 [features]

src/rust/include/rregex.h

@@ -5,3 +5,9 @@ struct Regex;
 extern struct Regex const * regex_new(char * const regex);
 extern uint64_t regex_matches(struct Regex const * const exp, uint8_t * const str, uint64_t str_len);
 extern void regex_free(struct Regex const * const exp);
+
+struct Regress;
+
+extern struct Regress const *regress_new(char * const regress);
+extern uint64_t regress_matches(struct Regress const * const exp, uint8_t * const str, uint64_t str_len);
+extern void regress_free(struct Regress const * const exp);

src/rust/src/lib.rs

@@ -1,7 +1,9 @@
 extern crate regex;
+extern crate regress;
 extern crate libc;
 
 use regex::bytes::Regex;
+use regress::Regex as Regress;
 use libc::c_char;
 use std::boxed::Box;
 use std::ffi::CStr;
@@ -34,3 +36,40 @@ pub extern fn regex_matches(raw_exp: *mut Regex, p: *const u8, len: u64) -> u64
 pub extern fn regex_free(raw_exp: *mut Regex) {
     unsafe { Box::from_raw(raw_exp) };
 }
+
+#[no_mangle]
+pub extern fn regress_new(c_buf: *const c_char) -> *const Regress {
+    let c_str: &CStr = unsafe { CStr::from_ptr(c_buf) };
+
+    // Pretend regress supports the "(?i)" syntax.
+    let mut pat = c_str.to_str().unwrap();
+    let mut flags = regress::Flags::default();
+    if pat.starts_with("(?i)") {
+        flags.icase = true;
+        pat = &pat["(?i)".len()..]
+    }
+    let exp = match regress::Regex::newf(pat, flags) {
+        Ok(val) => Box::into_raw(Box::new(val)),
+        Err(_) => ptr::null(),
+    };
+
+    exp as *const Regress
+}
+
+#[no_mangle]
+pub extern fn regress_matches(raw_exp: *mut Regress, p: *const u8, len: u64) -> u64 {
+    let exp = unsafe { Box::from_raw(raw_exp) };
+    let s = unsafe {
+        let sl = slice::from_raw_parts(p, len as usize);
+        std::str::from_utf8_unchecked(sl)
+    };
+
+    let findings = exp.find_iter_ascii(s).count();
+    Box::into_raw(exp);
+    findings as u64
+}
+
+#[no_mangle]
+pub extern fn regress_free(raw_exp: *mut Regress) {
+    unsafe { Box::from_raw(raw_exp) };
+}