Add support for regex function family

13 files changed, 623 insertions, 24 deletions

diff --git a/build2/buildfile b/build2/buildfile
index 8525eb8..d2aa7c1 100644
--- a/build2/buildfile
+++ b/build2/buildfile
@@ -19,6 +19,7 @@ exe{b}:                                                   \
             {            cxx}{ functions-filesystem     } \
             {            cxx}{ functions-path           } \
             {            cxx}{ functions-process-path   } \
+            {            cxx}{ functions-regex          } \
             {            cxx}{ functions-string         } \
             {            cxx}{ functions-target-triplet } \
             {hxx         cxx}{ lexer                    } \
diff --git a/build2/function.cxx b/build2/function.cxx
index 399d679..f7de749 100644
--- a/build2/function.cxx
+++ b/build2/function.cxx
@@ -304,6 +304,7 @@ namespace build2
   void filesystem_functions ();     // functions-filesystem.cxx
   void path_functions ();           // functions-path.cxx
   void process_path_functions ();   // functions-process-path.cxx
+  void regex_functions ();          // functions-regex.cxx
   void string_functions ();         // functions-string.cxx
   void target_triplet_functions (); // functions-target-triplet.cxx
 
@@ -315,6 +316,7 @@ namespace build2
       filesystem_functions ();
       path_functions ();
       process_path_functions ();
+      regex_functions ();
       string_functions ();
       target_triplet_functions ();
     }
diff --git a/build2/functions-regex.cxx b/build2/functions-regex.cxx
new file mode 100644
index 0000000..bdecc14
--- /dev/null
+++ b/build2/functions-regex.cxx
@@ -0,0 +1,331 @@
+// file      : build2/functions-regex.cxx -*- C++ -*-
+// copyright : Copyright (c) 2014-2017 Code Synthesis Ltd
+// license   : MIT; see accompanying LICENSE file
+
+#include <sstream>
+
+#include <libbutl/regex.hxx>
+
+#include <build2/function.hxx>
+#include <build2/variable.hxx>
+
+using namespace std;
+using namespace butl;
+
+namespace build2
+{
+  // Convert value of an arbitrary type to string.
+  //
+  static inline string
+  to_string (value&& v)
+  {
+    // Optimize for the string value type.
+    //
+    if (v.type != &value_traits<string>::value_type)
+      untypify (v);
+
+    return convert<string> (move (v));
+  }
+
+  // Parse a regular expression. Throw invalid_argument if it is not valid.
+  //
+  static regex
+  parse_regex (const string& s, regex::flag_type f)
+  {
+    try
+    {
+      return regex (s, f);
+    }
+    catch (const regex_error& e)
+    {
+      // Print regex_error description if meaningful (no space).
+      //
+      ostringstream os;
+      os << "invalid regex '" << s << "'" << e;
+      throw invalid_argument (os.str ());
+    }
+  }
+
+  // Match value of an arbitrary type against the regular expression. See
+  // match() overloads (below) for details.
+  //
+  static value
+  match (value&& v, const string& re, optional<names>&& flags)
+  {
+    // Parse flags.
+    //
+    regex::flag_type rf (regex::ECMAScript);
+    bool subs (false);
+
+    if (flags)
+    {
+      for (auto& f: *flags)
+      {
+        string s (convert<string> (move (f)));
+
+        if (s == "icase")
+          rf |= regex::icase;
+        else if (s == "return_subs")
+          subs = true;
+        else
+          throw invalid_argument ("invalid flag '" + s + "'");
+      }
+    }
+
+    // Parse regex.
+    //
+    regex rge (parse_regex (re, rf));
+
+    // Match.
+    //
+    string s (to_string (move (v)));
+
+    if (!subs)
+      return value (regex_match (s, rge)); // Return boolean value.
+
+    names r;
+    match_results<string::const_iterator> m;
+
+    if (regex_match (s, m, rge))
+    {
+      assert (!m.empty ());
+
+      for (size_t i (1); i != m.size (); ++i)
+      {
+        if (m[i].matched)
+          r.emplace_back (m.str (i));
+      }
+    }
+
+    return value (r);
+  }
+
+  // Determine if there is a match between the regular expression and some
+  // part of a value of an arbitrary type. See search() overloads (below)
+  // for details.
+  //
+  static value
+  search (value&& v, const string& re, optional<names>&& flags)
+  {
+    // Parse flags.
+    //
+    regex::flag_type rf (regex::ECMAScript);
+    bool match (false);
+    bool subs (false);
+
+    if (flags)
+    {
+      for (auto& f: *flags)
+      {
+        string s (convert<string> (move (f)));
+
+        if (s == "icase")
+          rf |= regex::icase;
+        else if (s == "return_match")
+          match = true;
+        else if (s == "return_subs")
+          subs = true;
+        else
+          throw invalid_argument ("invalid flag '" + s + "'");
+      }
+    }
+
+    // Parse regex.
+    //
+    regex rge (parse_regex (re, rf));
+
+    // Search.
+    //
+    string s (to_string (move (v)));
+
+    if (!match && !subs)
+      return value (regex_search (s, rge)); // Return boolean value.
+
+    names r;
+    match_results<string::const_iterator> m;
+
+    if (regex_search (s, m, rge))
+    {
+      assert (!m.empty ());
+
+      if (match)
+      {
+        assert (m[0].matched);
+        r.emplace_back (m.str (0));
+      }
+
+      if (subs)
+      {
+        for (size_t i (1); i != m.size (); ++i)
+        {
+          if (m[i].matched)
+            r.emplace_back (m.str (i));
+        }
+      }
+    }
+
+    return value (r);
+  }
+
+  // Replace matched parts in a value of an arbitrary type, using the format
+  // string. See replace() overloads (below) for details.
+  //
+  static names
+  replace (value&& v,
+           const string& re,
+           const string& fmt,
+           optional<names>&& flags)
+  {
+    // Parse flags.
+    //
+    regex::flag_type rf (regex::ECMAScript);
+    regex_constants::match_flag_type mf (regex_constants::match_default);
+
+    if (flags)
+    {
+      for (auto& f: *flags)
+      {
+        string s (convert<string> (move (f)));
+
+        if (s == "icase")
+          rf |= regex::icase;
+        else if (s == "format_first_only")
+          mf |= regex_constants::format_first_only;
+        else if (s == "format_no_copy")
+          mf |= regex_constants::format_no_copy;
+        else
+          throw invalid_argument ("invalid flag '" + s + "'");
+      }
+    }
+
+    // Parse regex.
+    //
+    regex rge (parse_regex (re, rf));
+
+    // Replace.
+    //
+    names r;
+
+    try
+    {
+      string s (to_string (move (v)));
+      r.emplace_back (regex_replace_ex (s, rge, fmt, mf).first);
+    }
+    catch (const regex_error& e)
+    {
+      fail << "unable to replace" << e;
+    }
+
+    return r;
+  }
+
+  void
+  regex_functions ()
+  {
+    function_family f ("regex");
+
+    // match
+    //
+    // Match a value of an arbitrary type against the regular expression.
+    // Convert the value to string prior to matching. Return the boolean value
+    // unless return_subs flag is specified (see below), in which case return
+    // names (empty if no match).
+    //
+    // The following flags are supported:
+    //
+    // icase       - match ignoring case
+    //
+    // return_subs - return names (rather than boolean), that contain
+    //               sub-strings that match the marked sub-expressions
+    //
+    f[".match"] = [](value s, string re, optional<names> flags)
+    {
+      return match (move (s), re, move (flags));
+    };
+
+    f[".match"] = [](value s, names re, optional<names> flags)
+    {
+      return match (move (s), convert<string> (move (re)), move (flags));
+    };
+
+    // search
+    //
+    // Determine if there is a match between the regular expression and some
+    // part of a value of an arbitrary type. Convert the value to string prior
+    // to searching. Return the boolean value unless return_match or
+    // return_subs flag is specified (see below) in which case return names
+    // (empty if no match).
+    //
+    // The following flags are supported:
+    //
+    // icase        - match ignoring case
+    //
+    // return_match - return names (rather than boolean), that contain a
+    //                sub-string that matches the whole regular expression
+    //
+    // return_subs -  return names (rather than boolean), that contain
+    //                sub-strings that match the marked sub-expressions
+    //
+    // If both return_match and return_subs flags are specified then the
+    // sub-string that matches the whole regular expression comes first.
+    //
+    f[".search"] = [](value s, string re, optional<names> flags)
+    {
+      return search (move (s), re, move (flags));
+    };
+
+    f[".search"] = [](value s, names re, optional<names> flags)
+    {
+      return search (move (s), convert<string> (move (re)), move (flags));
+    };
+
+    // replace
+    //
+    // Replace matched parts in a value of an arbitrary type, using the format
+    // string. Convert the value to string prior to matching. The result value
+    // is always untyped, regardless of the argument type.
+    //
+    // Substitution escape sequences are extended with a subset of Perl
+    // sequences (see regex_replace_ex() for details).
+    //
+    // The following flags are supported:
+    //
+    // icase             - match ignoring case
+    //
+    // format_first_only - only replace the first match
+    //
+    // format_no_copy    - do not copy unmatched value parts to the result
+    //
+    // If both format_first_only and format_no_copy flags are specified then
+    // all the result will contain is the replacement of the first match.
+    //
+    f[".replace"] = [](value s, string re, string fmt, optional<names> flags)
+    {
+      return replace (move (s), re, fmt, move (flags));
+    };
+
+    f[".replace"] = [](value s, string re, names fmt, optional<names> flags)
+    {
+      return replace (move (s),
+                      re,
+                      convert<string> (move (fmt)),
+                      move (flags));
+    };
+
+    f[".replace"] = [](value s, names re, string fmt, optional<names> flags)
+    {
+      return replace (move (s),
+                      convert<string> (move (re)),
+                      fmt,
+                      move (flags));
+    };
+
+    f[".replace"] = [](value s, names re, names fmt, optional<names> flags)
+    {
+      return replace (move (s),
+                      convert<string> (move (re)),
+                      convert<string> (move (fmt)),
+                      move (flags));
+    };
+  }
+}
diff --git a/build2/variable.hxx b/build2/variable.hxx
index 19a6c69..f0218fe 100644
--- a/build2/variable.hxx
+++ b/build2/variable.hxx
@@ -263,7 +263,7 @@ namespace build2
   bool operator>= (const value&, const value&);
 
   // Value cast. The first three expect the value to be not NULL. The cast
-  // from lookup expects the value to aslo be defined.
+  // from lookup expects the value to also be defined.
   //
   // Note that a cast to names expects the value to be untyped while a cast
   // to vector<names> -- typed.
diff --git a/tests/function/regex/buildfile b/tests/function/regex/buildfile
new file mode 100644
index 0000000..fbac223
--- /dev/null
+++ b/tests/function/regex/buildfile
@@ -0,0 +1,5 @@
+# file      : tests/function/regex/buildfile
+# copyright : Copyright (c) 2014-2017 Code Synthesis Ltd
+# license   : MIT; see accompanying LICENSE file
+
+./: test{testscript} $b
diff --git a/tests/function/regex/testscript b/tests/function/regex/testscript
new file mode 100644
index 0000000..deeefa5
--- /dev/null
+++ b/tests/function/regex/testscript
@@ -0,0 +1,255 @@
+# file      : tests/function/regex/testscript
+# copyright : Copyright (c) 2014-2017 Code Synthesis Ltd
+# license   : MIT; see accompanying LICENSE file
+
+.include ../../common.test
+
+: replace
+:
+{
+  : arg-types
+  :
+  {
+    : untyped-string-string
+    :
+    $* <<EOI >'foo.o'
+    print $regex.replace('foo.cxx', [string] '(^[^.]*).*', [string] '\1.o')
+    EOI
+
+    : string-untyped-string
+    :
+    $* <<EOI >'foo.o'
+    print $regex.replace([string] 'foo.cxx', '(^[^.]*).*', [string] '\1.o')
+    EOI
+
+    : bool-string-untyped
+    :
+    $* <<EOI >'true.o'
+    print $regex.replace('a' == "a", [string] '(^[^.]*).*', '\1.o')
+    EOI
+
+    : uint64-untyped-string
+    :
+    $* <<EOI >'1.o'
+    print $regex.replace([uint64] 01, '(^[^.]*).*', [string] '\1.o')
+    EOI
+
+    : path-untyped-untyped
+    :
+    $* <<EOI >'foo.o'
+    print $regex.replace([path] 'foo.cxx', '(^[^.]*).*', '\1.o')
+    EOI
+
+    : multiple-names
+    :
+    $* <<EOI 2>'error: invalid argument: invalid string value: multiple names' != 0
+    print $regex.replace(foo.cxx bar.cxx, '([^.]*)', '\1.o')
+    EOI
+
+    : null
+    :
+    $* <<EOI 2>'error: invalid argument: null value' != 0
+    print $regex.replace([null], '([^.]*)', '\1.o')
+    EOI
+
+    : null-regex
+    :
+    $* <<EOI 2>'error: invalid argument: null value' != 0
+    print $regex.replace(foo.cxx, [null], '\1.o')
+    EOI
+  }
+
+  : no-subs
+  :
+  $* <<EOI >'xbcxbc'
+  print $regex.replace('abcabc', 'a', 'x')
+  EOI
+
+  : no-match
+  :
+  $* <<EOI >'abcabc'
+  print $regex.replace('abcabc', 'd', 'x')
+  EOI
+
+  : flags
+  :
+  {
+    : icase
+    :
+    $* <<EOI >'Foo.o'
+    print $regex.replace("Foo.cxx", '(f[^.]*).*', '\1.o', icase)
+    EOI
+
+    : format_first-only
+    :
+    $* <<EOI >'foo.o'
+    print $regex.replace('foo.cxx', '([^.]*).*', '\1.o', format_first_only)
+    EOI
+
+    : format_no_copy
+    :
+    {
+      : all-matches
+      :
+      $* <<EOI >'xx'
+      print $regex.replace('abcabc', 'a', 'x', format_no_copy)
+      EOI
+
+      : first-only
+      :
+      $* <<EOI >'x'
+      print $regex.replace('abcabc', 'a', 'x', format_no_copy format_first_only)
+      EOI
+    }
+
+    : unknown
+    :
+    $* <<EOI 2>"error: invalid argument: invalid flag 'unknown'" != 0
+    print $regex.replace("foo.cxx", '(f[^.]*).*', '\1.o', unknown)
+    EOI
+  }
+
+  : invalid-regex
+  :
+  $* <'print $regex.replace(a, "[", b)' 2>>~/EOE/ != 0
+  /error: invalid argument: invalid regex '\['.*/
+  EOE
+}
+
+: match
+:
+{
+  : arg-types
+  :
+  {
+    : untyped-string
+    :
+    $* <<EOI >'true'
+    print $regex.match('foo.cxx', [string] '(^[^.]*).*')
+    EOI
+
+    : untyped-untyped
+    :
+    $* <<EOI >'true'
+    print $regex.match('foo.cxx', '(^[^.]*).*')
+    EOI
+  }
+
+  : flags
+  :
+  {
+    : none
+    :
+    $* <<EOI >'false'
+    print $regex.match("Foo.cxx", '(f[^.]*).*')
+    EOI
+
+    : icase
+    :
+    $* <<EOI >'true'
+    print $regex.match("Foo.cxx", '(f[^.]*).*', icase)
+    EOI
+
+    : return_subs
+    :
+    {
+      : success
+      :
+      $* <<EOI >'foo bar'
+      print $regex.match("foo   bar", '([^\s]*)\s+([^\s]*)', return_subs)
+      EOI
+
+      : no-subexpr
+      :
+      $* <<EOI >''
+      print $regex.match("foo   bar", '(?:[^\s]*)\s+(?:[^\s]*)', return_subs)
+      EOI
+
+      : failure
+      :
+      $* <<EOI >''
+      print $regex.match("   bar", '([^\s]+)\s+([^\s]+)', return_subs)
+      EOI
+    }
+  }
+}
+
+: search
+:
+{
+  : arg-types
+  :
+  {
+    : untyped-string
+    :
+    $* <<EOI >'true'
+    print $regex.search('.foo.cxx', [string] '([^.]*)')
+    EOI
+
+    : untyped-untyped
+    :
+    $* <<EOI >'true'
+    print $regex.search('.foo.cxx', '([^.]*)')
+    EOI
+  }
+
+  : flags
+  :
+  {
+    : none
+    :
+    $* <<EOI >'false'
+    print $regex.match("Foo.cxx", '(f[^.]*).*')
+    EOI
+
+    : icase
+    :
+    $* <<EOI >'true'
+    print $regex.match("Foo.cxx", '(f[^.]*).*', icase)
+    EOI
+
+    : return_subs
+    :
+    {
+      : success
+      :
+      $* <<EOI >'foo bar'
+      print $regex.search(" foo  bar baz", '([^\s]+)\s+([^\s]+)', return_subs)
+      EOI
+
+      : no-subexpr
+      :
+      $* <<EOI >''
+      print $regex.search("foo bar ba", '(?:[^\s]+)\s+(?:[^\s]+)', return_subs)
+      EOI
+
+      : failure
+      :
+      $* <<EOI >''
+      print $regex.match("   bar", '([^\s]+)\s+([^\s]+)', return_subs)
+      EOI
+    }
+
+    : return_match
+    :
+    {
+      : success
+      :
+      $* <<EOI >'foo  bar'
+      print $regex.search(" foo  bar baz", '([^\s]+)\s+([^\s]+)', return_match)
+      EOI
+
+      : subs
+      :
+      $* <<EOI >'foo  bar foo bar'
+      print $regex.search(" foo  bar baz", '([^\s]+)\s+([^\s]+)', return_match return_subs)
+      EOI
+
+      : failure
+      :
+      $* <<EOI >''
+      print $regex.search("   bar", '([^\s]+)\s+([^\s]+)', return_match)
+      EOI
+    }
+  }
+}
diff --git a/unit-tests/cc/lexer/buildfile b/unit-tests/cc/lexer/buildfile
index 3152e77..53adc33 100644
--- a/unit-tests/cc/lexer/buildfile
+++ b/unit-tests/cc/lexer/buildfile
@@ -5,11 +5,12 @@
 #@@ Temporary until we get utility library support.
 #
 import libs = libbutl%lib{butl}
-src = cc/lexer token lexer diagnostics utility variable name b-options \
+src = cc/lexer token lexer diagnostics utility variable name b-options     \
 types-parsers context scope parser target operation rule prerequisite file \
-module function functions-builtin functions-filesystem functions-path \
-functions-process-path functions-string functions-target-triplet algorithm \
-search dump filesystem scheduler config/{utility init operation module} spec
+module function functions-builtin functions-filesystem functions-path      \
+functions-process-path functions-regex functions-string                    \
+functions-target-triplet algorithm search dump filesystem scheduler        \
+config/{utility init operation module} spec
 
 exe{driver}: cxx{driver} ../../../build2/cxx{$src} ../../../build2/liba{b} \
 $libs test{*}
diff --git a/unit-tests/cc/parser/buildfile b/unit-tests/cc/parser/buildfile
index bb1ad53..59ef4fa 100644
--- a/unit-tests/cc/parser/buildfile
+++ b/unit-tests/cc/parser/buildfile
@@ -8,7 +8,7 @@ import libs = libbutl%lib{butl}
 src = cc/{lexer parser} token lexer diagnostics utility variable name    \
 b-options types-parsers context scope parser target operation rule       \
 prerequisite file module function functions-builtin functions-filesystem \
-functions-path functions-process-path functions-string                   \
+functions-path functions-process-path functions-regex functions-string   \
 functions-target-triplet algorithm search dump filesystem scheduler      \
 config/{utility init operation module} spec
 
diff --git a/unit-tests/function/buildfile b/unit-tests/function/buildfile
index c0cd26f..ad29ac7 100644
--- a/unit-tests/function/buildfile
+++ b/unit-tests/function/buildfile
@@ -5,11 +5,11 @@
 #@@ Temporary until we get utility library support.
 #
 import libs = libbutl%lib{butl}
-src = token lexer diagnostics utility variable name b-options types-parsers \
+src = token lexer diagnostics utility variable name b-options types-parsers  \
 context scope parser target operation rule prerequisite file module function \
 functions-builtin functions-filesystem functions-path functions-process-path \
-functions-string functions-target-triplet algorithm search dump filesystem \
-scheduler config/{utility init operation module} spec
+functions-regex functions-string functions-target-triplet algorithm search   \
+dump filesystem scheduler config/{utility init operation module} spec
 
 exe{driver}: cxx{driver} ../../build2/cxx{$src} ../../build2/liba{b} $libs test{call syntax}
 
diff --git a/unit-tests/lexer/buildfile b/unit-tests/lexer/buildfile
index 3d4e3ed..0ab50de 100644
--- a/unit-tests/lexer/buildfile
+++ b/unit-tests/lexer/buildfile
@@ -5,11 +5,11 @@
 #@@ Temporary until we get utility library support.
 #
 import libs = libbutl%lib{butl}
-src = token lexer diagnostics utility variable name b-options types-parsers \
+src = token lexer diagnostics utility variable name b-options types-parsers  \
 context scope parser target operation rule prerequisite file module function \
 functions-builtin functions-filesystem functions-path functions-process-path \
-functions-string functions-target-triplet algorithm search dump filesystem \
-scheduler config/{utility init operation module} spec
+functions-regex functions-string functions-target-triplet algorithm search   \
+dump filesystem scheduler config/{utility init operation module} spec
 
 exe{driver}: cxx{driver} ../../build2/cxx{$src} ../../build2/liba{b} $libs \
 test{*}
diff --git a/unit-tests/scheduler/buildfile b/unit-tests/scheduler/buildfile
index 22361b5..b3b97ce 100644
--- a/unit-tests/scheduler/buildfile
+++ b/unit-tests/scheduler/buildfile
@@ -5,11 +5,11 @@
 #@@ Temporary until we get utility library support.
 #
 import libs = libbutl%lib{butl}
-src = token lexer diagnostics utility variable name b-options types-parsers \
+src = token lexer diagnostics utility variable name b-options types-parsers  \
 context scope parser target operation rule prerequisite file module function \
 functions-builtin functions-filesystem functions-path functions-process-path \
-functions-string functions-target-triplet algorithm search dump filesystem \
-scheduler config/{utility init operation module} spec
+functions-regex functions-string functions-target-triplet algorithm search   \
+dump filesystem scheduler config/{utility init operation module} spec
 
 exe{driver}: cxx{driver} ../../build2/cxx{$src} ../../build2/liba{b} $libs
 
diff --git a/unit-tests/test/script/lexer/buildfile b/unit-tests/test/script/lexer/buildfile
index 2070089..a21df7a 100644
--- a/unit-tests/test/script/lexer/buildfile
+++ b/unit-tests/test/script/lexer/buildfile
@@ -5,13 +5,15 @@
 #@@ Temporary until we get utility library support.
 #
 import libs = libbutl%lib{butl}
-src = token lexer diagnostics utility variable name b-options types-parsers \
+src = token lexer diagnostics utility variable name b-options types-parsers  \
 context scope parser target operation rule prerequisite file module function \
 functions-builtin functions-filesystem functions-path functions-process-path \
-functions-string functions-target-triplet algorithm search dump filesystem \
-scheduler config/{utility init operation module} test/script/{token lexer} spec
+functions-regex functions-string functions-target-triplet algorithm search   \
+dump filesystem scheduler config/{utility init operation module}             \
+test/script/{token lexer} spec
 
-exe{driver}: cxx{driver} ../../../../build2/cxx{$src} ../../../../build2/liba{b} $libs \
+exe{driver}: cxx{driver} ../../../../build2/cxx{$src}                      \
+../../../../build2/liba{b} $libs                                           \
 test{command-line first-token second-token command-expansion variable-line \
      description-line variable}
 
diff --git a/unit-tests/test/script/parser/buildfile b/unit-tests/test/script/parser/buildfile
index 634120c..184fd82 100644
--- a/unit-tests/test/script/parser/buildfile
+++ b/unit-tests/test/script/parser/buildfile
@@ -8,12 +8,14 @@ import libs = libbutl%lib{butl}
 src = token lexer parser diagnostics utility variable name context target \
 scope prerequisite file module operation rule b-options algorithm search  \
 filesystem function functions-builtin functions-filesystem functions-path \
-functions-process-path functions-string functions-target-triplet          \
-config/{utility init operation module} dump types-parsers                 \
-test/{target script/{token lexer parser regex script}} scheduler spec
+functions-process-path functions-regex functions-string                   \
+functions-target-triplet config/{utility init operation module} dump      \
+types-parsers test/{target script/{token lexer parser regex script}}      \
+scheduler spec
 
-exe{driver}: cxx{driver} ../../../../build2/cxx{$src} ../../../../build2/liba{b} $libs \
-test{cleanup command-if command-re-parse description directive exit \
+exe{driver}: cxx{driver} ../../../../build2/cxx{$src}                \
+../../../../build2/liba{b} $libs                                     \
+test{cleanup command-if command-re-parse description directive exit  \
      expansion here-document here-string include pipe-expr pre-parse \
      redirect regex scope scope-if setup-teardown}