[ASTMatchers] Enhanced support for matchers taking Regex arguments

Added new Macros `AST(_POLYMORPHIC)_MATCHER_REGEX(_OVERLOAD)` that define a matchers that take a regular expression string and optionally regular expression flags. This lets users match against nodes while ignoring the case without having to manually use `[Aa]` or `[A-Fa-f]` in their regex. The other point this addresses is in the current state, matchers that use regular expressions have to compile them for each node they try to match on, Now the regular expression is compiled once when you define the matcher and used for every node that it tries to match against. If there is an error while compiling the regular expression an error will be logged to stderr showing the bad regex string and the reason it couldn't be compiled. The old behaviour of this was down to the Matcher implementation and some would assert, whereas others just would never match. Support for this has been added to the documentation script as well. Support for this has been added to dynamic matchers ensuring functionality is the same between the 2 use cases.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D82706

include/llvm/Support/Regex.h

@@ -16,6 +16,7 @@
 #ifndef LLVM_SUPPORT_REGEX_H
 #define LLVM_SUPPORT_REGEX_H
 
+#include "llvm/ADT/BitmaskEnum.h"
 #include <string>
 
 struct llvm_regex;
@@ -26,20 +27,22 @@ namespace llvm {
 
   class Regex {
   public:
-    enum {
-      NoFlags=0,
+    enum RegexFlags : unsigned {
+      NoFlags = 0,
       /// Compile for matching that ignores upper/lower case distinctions.
-      IgnoreCase=1,
+      IgnoreCase = 1,
       /// Compile for newline-sensitive matching. With this flag '[^' bracket
       /// expressions and '.' never match newline. A ^ anchor matches the
       /// null string after any newline in the string in addition to its normal
       /// function, and the $ anchor matches the null string before any
       /// newline in the string in addition to its normal function.
-      Newline=2,
+      Newline = 2,
       /// By default, the POSIX extended regular expression (ERE) syntax is
       /// assumed. Pass this flag to turn on basic regular expressions (BRE)
       /// instead.
-      BasicRegex=4
+      BasicRegex = 4,
+
+      LLVM_MARK_AS_BITMASK_ENUM(BasicRegex)
     };
 
     Regex();
@@ -47,7 +50,8 @@ namespace llvm {
     ///
     /// \param Regex - referenced string is no longer needed after this
     /// constructor does finish.  Only its compiled form is kept stored.
-    Regex(StringRef Regex, unsigned Flags = NoFlags);
+    Regex(StringRef Regex, RegexFlags Flags = NoFlags);
+    Regex(StringRef Regex, unsigned Flags);
     Regex(const Regex &) = delete;
     Regex &operator=(Regex regex) {
       std::swap(preg, regex.preg);

lib/Support/Regex.cpp

@@ -26,7 +26,7 @@ using namespace llvm;
 
 Regex::Regex() : preg(nullptr), error(REG_BADPAT) {}
 
-Regex::Regex(StringRef regex, unsigned Flags) {
+Regex::Regex(StringRef regex, RegexFlags Flags) {
   unsigned flags = 0;
   preg = new llvm_regex();
   preg->re_endp = regex.end();
@@ -39,6 +39,9 @@ Regex::Regex(StringRef regex, unsigned Flags) {
   error = llvm_regcomp(preg, regex.data(), flags|REG_PEND);
 }
 
+Regex::Regex(StringRef regex, unsigned Flags)
+    : Regex(regex, static_cast<RegexFlags>(Flags)) {}
+
 Regex::Regex(Regex &&regex) {
   preg = regex.preg;
   error = regex.error;