Convert slice to use Regex (#12570)

* Convert slice to use Regex

* swap conditions

* add comments for regex execs

Author: cmcfarlen

plugins/slice/Config.cc

@@ -33,15 +33,8 @@ constexpr std::string_view DefaultCrrIdentHeader  = {"X-Crr-Ident"};
 
 Config::~Config()
 {
-  if (nullptr != m_regex_extra) {
-#ifndef PCRE_STUDY_JIT_COMPILE
-    pcre_free(m_regex_extra);
-#else
-    pcre_free_study(m_regex_extra);
-#endif
-  }
   if (nullptr != m_regex) {
-    pcre_free(m_regex);
+    delete m_regex;
   }
 }
 
@@ -159,16 +152,16 @@ Config::fromArgs(int const argc, char const *const argv[])
         break;
       }
 
-      const char *errptr;
+      std::string err;
       int         erroffset;
       m_regexstr = optarg;
-      m_regex    = pcre_compile(m_regexstr.c_str(), 0, &errptr, &erroffset, nullptr);
-      if (nullptr == m_regex) {
-        ERROR_LOG("Invalid regex: '%s'", m_regexstr.c_str());
-      } else {
-        m_regex_type  = Exclude;
-        m_regex_extra = pcre_study(m_regex, 0, &errptr);
+      m_regex    = new Regex();
+
+      if (m_regex->compile(m_regexstr, err, erroffset)) {
+        m_regex_type = Exclude;
         DEBUG_LOG("Using regex for url exclude: '%s'", m_regexstr.c_str());
+      } else {
+        ERROR_LOG("Invalid regex: '%s' - %s at column %d", m_regexstr.c_str(), err.c_str(), erroffset);
       }
     } break;
     case 'g': {
@@ -180,17 +173,16 @@ Config::fromArgs(int const argc, char const *const argv[])
         ERROR_LOG("Regex already specified!");
         break;
       }
-
-      const char *errptr;
+      std::string err;
       int         erroffset;
       m_regexstr = optarg;
-      m_regex    = pcre_compile(m_regexstr.c_str(), 0, &errptr, &erroffset, nullptr);
-      if (nullptr == m_regex) {
-        ERROR_LOG("Invalid regex: '%s'", m_regexstr.c_str());
-      } else {
-        m_regex_type  = Include;
-        m_regex_extra = pcre_study(m_regex, 0, &errptr);
+      m_regex    = new Regex();
+
+      if (m_regex->compile(m_regexstr, err, erroffset)) {
+        m_regex_type = Include;
         DEBUG_LOG("Using regex for url include: '%s'", m_regexstr.c_str());
+      } else {
+        ERROR_LOG("Invalid regex: '%s' - %s at column %d", m_regexstr.c_str(), err.c_str(), erroffset);
       }
     } break;
     case 'l': {
@@ -330,12 +322,14 @@ Config::matchesRegex(char const *const url, int const urllen) const
 
   switch (m_regex_type) {
   case Exclude: {
-    if (0 <= pcre_exec(m_regex, m_regex_extra, url, urllen, 0, 0, nullptr, 0)) {
+    // Exclude means if the regex matches, it doesn't match
+    if (m_regex->exec({url, static_cast<size_t>(urllen)})) {
       matches = false;
     }
   } break;
   case Include: {
-    if (pcre_exec(m_regex, m_regex_extra, url, urllen, 0, 0, nullptr, 0) < 0) {
+    // Include means if the regex matches, it matches
+    if (!m_regex->exec({url, static_cast<size_t>(urllen)})) {
       matches = false;
     }
   } break;

plugins/slice/Config.h

@@ -18,15 +18,11 @@
 
 #pragma once
 
+#include <tsutil/Regex.h>
+
 #include "slice.h"
 #include "ObjectSizeCache.h"
 
-#ifdef HAVE_PCRE_PCRE_H
-#include <pcre/pcre.h>
-#else
-#include <pcre.h>
-#endif
-
 #include <string>
 #include <mutex>
 
@@ -40,11 +36,10 @@ struct Config {
   std::string m_remaphost; // remap host to use for loopback slice GET
   std::string m_regexstr;  // regex string for things to slice (default all)
   enum RegexType { None, Include, Exclude };
-  RegexType   m_regex_type{None};
-  pcre       *m_regex{nullptr};
-  pcre_extra *m_regex_extra{nullptr};
-  int         m_paceerrsecs{0};   // -1 disable logging, 0 no pacing, max 60s
-  int         m_prefetchcount{0}; // 0 disables prefetching
+  RegexType m_regex_type{None};
+  Regex    *m_regex{nullptr};
+  int       m_paceerrsecs{0};   // -1 disable logging, 0 no pacing, max 60s
+  int       m_prefetchcount{0}; // 0 disables prefetching
   enum RefType { First, Relative };
   RefType  m_reftype{First};          // reference slice is relative to request
   bool     m_head_strip_range{false}; // strip range header for head requests