regex: Allow to configure global/default options for regexes.

Like caseless matching or extended format, which may be useful as default
behaviour the whole configuration.



git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@1824339 13f79535-47bb-0310-9956-ffa450edef68

CHANGES

@@ -1,6 +1,9 @@
                                                          -*- coding: utf-8 -*-
 Changes with Apache 2.5.1
 
+  *) regex: Allow to configure global/default options for regexes, like
+     caseless matching or extended format.  [Yann Ylavic]
+
   *) mod_authnz_ldap: Fix language long names detection as short name.
      [Yann Ylavic]
 

include/ap_regex.h

@@ -80,6 +80,8 @@ extern "C" {
 #define AP_REG_NOTEMPTY 0x080 /**< Empty match not valid */
 #define AP_REG_ANCHORED 0x100 /**< Match at the first position */
 
+#define AP_REG_DOLLAR_ENDONLY 0x200 /**< '$' matches at end of subject string only */
+
 #define AP_REG_MATCH "MATCH_" /**< suggested prefix for ap_regname */
 
 /* Arguments for ap_pcre_version_string */
@@ -120,6 +122,26 @@ typedef struct {
  */
 AP_DECLARE(const char *) ap_pcre_version_string(int which);
 
+/**
+ * Get default compile flags
+ * @return Bitwise OR of AP_REG_* flags
+ */
+AP_DECLARE(int) ap_regcomp_get_default_cflags(void);
+
+/**
+ * Set default compile flags
+ * @param cflags Bitwise OR of AP_REG_* flags
+ */
+AP_DECLARE(void) ap_regcomp_set_default_cflags(int cflags);
+
+/**
+ * Get the AP_REG_* corresponding to the string.
+ * @param name The name (i.e. AP_REG_<name>)
+ * @return The AP_REG_*, or zero if the string is unknown
+ *
+ */
+AP_DECLARE(int) ap_regcomp_default_cflag_by_name(const char *name);
+
 /**
  * Compile a regular expression.
  * @param preg Returned compiled regex

server/core.c

@@ -49,6 +49,7 @@
 #include "mod_proxy.h"
 #include "ap_listen.h"
 #include "ap_provider.h"
+#include "ap_regex.h"
 
 #include "mod_so.h" /* for ap_find_loaded_module_symbol */
 
@@ -2909,6 +2910,58 @@ static const char *virtualhost_section(cmd_parms *cmd, void *dummy,
     return errmsg;
 }
 
+static const char *set_regex_default_options(cmd_parms *cmd,
+                                             void *dummy,
+                                             const char *arg)
+{
+    const command_rec *thiscmd = cmd->cmd;
+    int cflags, cflag;
+
+    const char *err = ap_check_cmd_context(cmd, GLOBAL_ONLY);
+    if (err != NULL) {
+        return err;
+    }
+
+    cflags = ap_regcomp_get_default_cflags();
+    while (*arg) {
+        const char *name = ap_getword_conf(cmd->pool, &arg);
+        int how = 0;
+
+        if (strcasecmp(name, "none") == 0) {
+            cflags = 0;
+            continue;
+        }
+
+        if (*name == '+') {
+            name++;
+            how = +1;
+        }
+        else if (*name == '-') {
+            name++;
+            how = -1;
+        }
+
+        cflag = ap_regcomp_default_cflag_by_name(name);
+        if (!cflag) {
+            return apr_psprintf(cmd->pool, "%s: option '%s' unknown",
+                                thiscmd->name, name);
+        }
+
+        if (how > 0) {
+            cflags |= cflag;
+        }
+        else if (how < 0) {
+            cflags &= ~cflag;
+        }
+        else {
+            cflags = cflag;
+        }
+    }
+    ap_regcomp_set_default_cflags(cflags);
+
+    return NULL;
+}
+
 static const char *set_server_alias(cmd_parms *cmd, void *dummy,
                                     const char *arg)
 {
@@ -4645,6 +4698,9 @@ AP_INIT_TAKE12("RLimitNPROC", no_set_limit, NULL,
    OR_ALL, "soft/hard limits for max number of processes per uid"),
 #endif
 
+AP_INIT_RAW_ARGS("RegexDefaultOptions", set_regex_default_options, NULL, RSRC_CONF,
+                 "default options for regexes (prefixed by '+' to add, '-' to del)"),
+
 /* internal recursion stopper */
 AP_INIT_TAKE12("LimitInternalRecursion", set_recursion_limit, NULL, RSRC_CONF,
               "maximum recursion depth of internal redirects and subrequests"),
@@ -5089,6 +5145,8 @@ static int core_pre_config(apr_pool_t *pconf, apr_pool_t *plog, apr_pool_t *ptem
     apr_pool_cleanup_register(pconf, NULL, reset_config_defines,
                               apr_pool_cleanup_null);
 
+    ap_regcomp_set_default_cflags(AP_REG_DOLLAR_ENDONLY);
+
     mpm_common_pre_config(pconf);
 
     return OK;

server/util_pcre.c

@@ -148,6 +148,38 @@ AP_DECLARE(void) ap_regfree(ap_regex_t *preg)
  *            Compile a regular expression       *
  *************************************************/
 
+static int default_cflags = AP_REG_DOLLAR_ENDONLY;
+
+AP_DECLARE(int) ap_regcomp_get_default_cflags(void)
+{
+    return default_cflags;
+}
+
+AP_DECLARE(void) ap_regcomp_set_default_cflags(int cflags)
+{
+    default_cflags = cflags;
+}
+
+AP_DECLARE(int) ap_regcomp_default_cflag_by_name(const char *name)
+{
+    int cflag = 0;
+
+    if (strcasecmp(name, "ICASE") == 0) {
+        cflag = AP_REG_ICASE;
+    }
+    else if (strcasecmp(name, "DOTALL") == 0) {
+        cflag = AP_REG_DOTALL;
+    }
+    else if (strcasecmp(name, "DOLLAR_ENDONLY") == 0) {
+        cflag = AP_REG_DOLLAR_ENDONLY;
+    }
+    else if (strcasecmp(name, "EXTENDED") == 0) {
+        cflag = AP_REG_EXTENDED;
+    }
+
+    return cflag;
+}
+
 /*
  * Arguments:
  *  preg        points to a structure for recording the compiled expression
@@ -169,12 +201,15 @@ AP_DECLARE(int) ap_regcomp(ap_regex_t * preg, const char *pattern, int cflags)
     int errcode = 0;
     int options = PCREn(DUPNAMES);
 
+    cflags |= default_cflags;
     if ((cflags & AP_REG_ICASE) != 0)
         options |= PCREn(CASELESS);
     if ((cflags & AP_REG_NEWLINE) != 0)
         options |= PCREn(MULTILINE);
     if ((cflags & AP_REG_DOTALL) != 0)
         options |= PCREn(DOTALL);
+    if ((cflags & AP_REG_DOLLAR_ENDONLY) != 0)
+        options |= PCREn(DOLLAR_ENDONLY);
 
 #ifdef HAVE_PCRE2
     preg->re_pcre = pcre2_compile((const unsigned char *)pattern,