diff --git a/doc/src/history.adoc b/doc/src/history.adoc
index 7997a0ca56..cf4b9892e9 100644
--- a/doc/src/history.adoc
+++ b/doc/src/history.adoc
@@ -6,6 +6,8 @@
 * Fix jam/{CPP} bind definitions of 4 or more values in a single
   declared argument not actually adding all the definitions.
   -- _Paolo Pastori_
+* *New*: Added test for regular expressions with MATCH builtin rule.
+  -- _Paolo Pastori_
 
 == Version 5.4.2
 
diff --git a/test/README.adoc b/test/README.adoc
index c407ee01c5..b4d0d245d8 100644
--- a/test/README.adoc
+++ b/test/README.adoc
@@ -311,7 +311,7 @@ time. All the elements in `names` should be relative paths.
 
 . Stores the state of the working directory in `self.previous_tree`.
 . Changes to `subdir`, if it is specified. It is relative to the
-  `original_workdir` or the workdir specified in `__init`.
+  `original_workdir` or the workdir specified in `+__init__+`.
 . Invokes the `b2` executable, passing `extra_args` to it. The binary should be
   located under `<test_invocation_dir>/../src/engine`. This is to make sure
   tests use the version of `b2` build from source.
@@ -376,7 +376,7 @@ The members are:
 * `expect_modification`
 * `expect_nothing`
 
-Note that `expect_modification` is used to check that a either file content or
+Note that `expect_modification` is used to check if the file contents or
 timestamp has changed. The rationale is that some compilers change content even
 if sources does not change, and it's easier to have a method which checks for
 both content and time changes.
diff --git a/test/match_list.py b/test/match_list.py
new file mode 100644
index 0000000000..2e9e0bb98f
--- /dev/null
+++ b/test/match_list.py
@@ -0,0 +1,94 @@
+#!/usr/bin/env python3
+
+# Copyright 2026 Paolo Pastori
+# Distributed under the Boost Software License, Version 1.0.
+# (See accompanying file LICENSE.txt or https://www.bfgroup.xyz/b2/LICENSE.txt)
+
+# List of tested regexps. Each tuple contains the pattern,
+# the value used for the test, and the expected result (do not use [] here).
+# Any test is processed by an instruction such as
+#
+# ECHO [ MATCH <pattern> : <value> ] ;
+#
+# fell free to add more test cases...
+
+# remember to use raw strings to avoid surprises with escaping
+trials = [
+    (r'', r'', r''),
+    # nothing captured (no parentheses in the pattern)
+    (r'"hello world"', r'"hello world"', r''),
+    # captures (more than one)
+    (r'(first(second))(third)', r'firstsecondthird', r'firstsecond second third'),
+    # ^ matches at start of line
+    (r'^(world)', r'"hello world"', r''),
+    (r'(world)', r'"hello world"', r'world'),
+    # $ matches at end of line
+    (r'(hello)$', r'"hello world"', r''),
+    # . matches any single character
+    (r'(.)', r'x', r'x'),
+    # literal dot (with escaping looses special meaning)
+    # NOTE: a \ followed by one of the characters ^.[$()|*+?\
+    # matches that character taken as an ordinary character,
+    # while a \ followed by any other character (but <>) does nothing!
+    # NOTE: because of common escaping by shell/interpreters
+    # to obtain a final \ you often have to escape itself using
+    # \\ or enclose it in raw strings (Python r'..', C++ R"...")
+    (r'(\\.)', r'y', r''),
+    (r'(\\.)', r'.', r'.'),
+    # ? matches an optional atom, matches a sequence
+    # of 0 or 1 matches of the atom
+    (r'bar(s)?', r'bar', r''),
+    (r'bar(s)?', r'bars', r's'),
+    # + matches a sequence of 1 or more matches of the atom
+    (r'(cin)+', r'cin', r'cin'),
+    (r'((cin)+)', r'cincin', r'cincin cin'),
+    # * matches a sequence of 0 or more matches of the atom
+    (r'(0)*', r'1', r''),
+    (r'(0)*', r'1000', r''),     # NOTE: this does not work as expected
+                                 #       at the beginning of the pattern
+    (r'1(0)*', r'1000', r'0'),
+    (r'1(0)*1$', r'1001', r'0'),
+    # \< matches at the beginning of a word
+    (r'\\<(lo)', r'hello', r''),
+    (r'\\<(lo)', r"she's so lovely", r'lo'),
+    # \> matches at the end of a word
+    (r'"\\>( fi)"', r'fidel', r''),
+    (r'"\\>( fi)"', r'"hi fi"', r' fi'), # NOTE: extra space in result too
+    # | separate branches, matches anything that matches one of the branches
+    (r'(left)|(right)', r'left', r'left'),
+    (r'(left)|(right)', r'right', r' right'), # NOTE: extra space as first group is empty
+    # [] character class (list of characters enclosed in []), matches
+    # any single character from the list. If the list begins with ^, it
+    # matches any single character not from the rest of the list.
+    # If two characters in the list are separated by -, this is shorthand
+    # for the full range of characters between those two.
+    # To include a literal ] in the list, make it the first character
+    # (following a possible ^). To include a literal -, make it the first
+    # or last character. Within brackets special characters ^.[$()|*+?
+    # loose their special meaning.
+    (r'"([0-9]+)"', r'1980s', r'1980'),
+    (r'"([^0-9]+)"', r'1980s', r's'),
+    # some real life cases
+    (r'"^([0-9]+)\\.([0-9]+)(.*)$"', r'5.4.3beta', r'5 4 .3beta'),
+    (r'^@(.*)', r'@my-rule', r'my-rule'),
+    (r'^(!)?(.*)', r'"!bla bla"', r'! bla bla'),
+]
+
+# Do not change code below !
+
+testln = []
+exptln = []
+for n, c in enumerate(trials):
+    testln.append('ECHO {} [ MATCH {} : {} ] ;'.format(n, c[0], c[1]))
+    exptln.append('{} {}'.format(n, c[2]) if c[2] else str(n))
+testln.append('EXIT : 0 ;\n')
+exptln.append('\n')
+
+import BoostBuild
+
+t = BoostBuild.Tester(pass_toolset=False)
+t.write('Jamroot', '\n'.join(testln))
+t.run_build_system()
+t.expect_output_lines('\n'.join(exptln))
+t.expect_nothing_more()
+t.cleanup()
diff --git a/test/test_all.py b/test/test_all.py
index 8bfa81ea9f..fff29a1d86 100755
--- a/test/test_all.py
+++ b/test/test_all.py
@@ -394,6 +394,7 @@ def reorder_tests(tests, first_test):
     "load_order",
     "loop",
     "make_rule",
+    "match_list",
     "message",
     "ndebug",
     "no_type",