enhance apply_regex_substitutions to support multi-line matches

It might be required to replace patterns with more context, e.g. content
of the next or previous line to disambiguate otherwise too generic matches.

Add parameter `single_line` to enable the old behavior (default) of
matching per line and otherwise match the whole text.
Add parameter `match_all` to require all patterns to match for each file
not only at least one.

Author: Flamefire

easybuild/tools/filetools.py

@@ -1643,16 +1643,22 @@ def apply_patch(patch_file, dest, fn=None, copy=False, level=None, use_git_am=Fa
     return True
 
 
-def apply_regex_substitutions(paths, regex_subs, backup='.orig.eb', on_missing_match=None):
+def apply_regex_substitutions(paths, regex_subs, backup='.orig.eb',
+                              on_missing_match=None, match_all=False, single_line=True):
     """
     Apply specified list of regex substitutions.
 
     :param paths: list of paths to files to patch (or just a single filepath)
-    :param regex_subs: list of substitutions to apply, specified as (<regexp pattern>, <replacement string>)
+    :param regex_subs: list of substitutions to apply,
+                       specified as (<regexp pattern or regex instance>, <replacement string>)
     :param backup: create backup of original file with specified suffix (no backup if value evaluates to False)
     :param on_missing_match: Define what to do when no match was found in the file.
                              Can be 'error' to raise an error, 'warn' to print a warning or 'ignore' to do nothing
                              Defaults to the value of --strict
+    :param match_all: Expect to match all patterns in all files
+                      instead of at least one per file for error/warning reporting
+    :param single_line: Replace first match of each pattern for each line in the order of the patterns.
+                        If False the patterns are applied in order to the full text and may match line breaks.
     """
     if on_missing_match is None:
         on_missing_match = build_option('strict')
@@ -1664,18 +1670,22 @@ def apply_regex_substitutions(paths, regex_subs, backup='.orig.eb', on_missing_m
     if isinstance(paths, string_type):
         paths = [paths]
 
+    flags = 0 if single_line else re.M
+    compiled_regex_subs = [(re.compile(regex, flags) if isinstance(regex, str) else regex, subtxt)
+                           for (regex, subtxt) in regex_subs]
+
     # only report when in 'dry run' mode
     if build_option('extended_dry_run'):
         paths_str = ', '.join(paths)
         dry_run_msg("applying regex substitutions to file(s): %s" % paths_str, silent=build_option('silent'))
-        for regex, subtxt in regex_subs:
-            dry_run_msg("  * regex pattern '%s', replacement string '%s'" % (regex, subtxt))
+        for regex, subtxt in compiled_regex_subs:
+            dry_run_msg("  * regex pattern '%s', replacement string '%s'" % (regex.pattern, subtxt))
 
     else:
-        _log.info("Applying following regex substitutions to %s: %s", paths, regex_subs)
-
-        compiled_regex_subs = [(re.compile(regex), subtxt) for (regex, subtxt) in regex_subs]
+        _log.info("Applying following regex substitutions to %s: %s",
+                  paths, [(regex.pattern, subtxt) for regex, subtxt in compiled_regex_subs])
 
+        replacement_failed_msgs = []
         for path in paths:
             try:
                 # make sure that file can be opened in text mode;
@@ -1695,32 +1705,49 @@ def apply_regex_substitutions(paths, regex_subs, backup='.orig.eb', on_missing_m
                 if backup:
                     copy_file(path, path + backup)
                 replacement_msgs = []
+                replaced = [False] * len(compiled_regex_subs)
                 with open_file(path, 'w') as out_file:
-                    lines = txt_utf8.split('\n')
-                    del txt_utf8
-                    for line_id, line in enumerate(lines):
-                        for regex, subtxt in compiled_regex_subs:
-                            match = regex.search(line)
-                            if match:
+                    if single_line:
+                        lines = txt_utf8.split('\n')
+                        del txt_utf8
+                        for line_id, line in enumerate(lines):
+                            for i, (regex, subtxt) in enumerate(compiled_regex_subs):
+                                match = regex.search(line)
+                                if match:
+                                    origtxt = match.group(0)
+                                    replacement_msgs.append("Replaced in line %d: '%s' -> '%s'" %
+                                                            (line_id + 1, origtxt, subtxt))
+                                    replaced[i] = True
+                                    line = regex.sub(subtxt, line)
+                                    lines[line_id] = line
+                        out_file.write('\n'.join(lines))
+                    else:
+                        for i, (regex, subtxt) in enumerate(compiled_regex_subs):
+                            def do_replace(match):
                                 origtxt = match.group(0)
-                                replacement_msgs.append("Replaced in line %d: '%s' -> '%s'" %
-                                                        (line_id + 1, origtxt, subtxt))
-                                line = regex.sub(subtxt, line)
-                                lines[line_id] = line
-                    out_file.write('\n'.join(lines))
+                                # pylint: disable=cell-var-from-loop
+                                cur_subtxt = match.expand(subtxt)
+                                # pylint: disable=cell-var-from-loop
+                                replacement_msgs.append("Replaced: '%s' -> '%s'" % (origtxt, cur_subtxt))
+                                return cur_subtxt
+                            txt_utf8, replaced[i] = regex.subn(do_replace, txt_utf8)
+                        out_file.write(txt_utf8)
                 if replacement_msgs:
                     _log.info('Applied the following substitutions to %s:\n%s', path, '\n'.join(replacement_msgs))
-                else:
-                    msg = 'Nothing found to replace in %s' % path
-                    if on_missing_match == ERROR:
-                        raise EasyBuildError(msg)
-                    elif on_missing_match == WARN:
-                        _log.warning(msg)
-                    else:
-                        _log.info(msg)
-
+                if (match_all and not all(replaced)) or (not match_all and not any(replaced)):
+                    errors = ["Nothing found to replace '%s'" % regex.pattern
+                              for cur_replaced, (regex, _) in zip(replaced, compiled_regex_subs) if not cur_replaced]
+                    replacement_failed_msgs.append(', '.join(errors) + ' in ' + path)
             except (IOError, OSError) as err:
                 raise EasyBuildError("Failed to patch %s: %s", path, err)
+            if replacement_failed_msgs:
+                msg = '\n'.join(replacement_failed_msgs)
+                if on_missing_match == ERROR:
+                    raise EasyBuildError(msg)
+                elif on_missing_match == WARN:
+                    _log.warning(msg)
+                else:
+                    _log.info(msg)
 
 
 def modify_env(old, new):

test/framework/filetools.py

@@ -1443,16 +1443,27 @@ def test_apply_regex_substitutions(self):
         # Check handling of on_missing_match
         ft.write_file(testfile, testtxt)
         regex_subs_no_match = [('Not there', 'Not used')]
-        error_pat = 'Nothing found to replace in %s' % testfile
+        error_pat = "Nothing found to replace 'Not there' in %s" % testfile
         # Error
         self.assertErrorRegex(EasyBuildError, error_pat, ft.apply_regex_substitutions, testfile, regex_subs_no_match,
                               on_missing_match=run.ERROR)
+        # First matches, but 2nd not
+        regex_subs_part_match = [regex_subs[0], ('Not there', 'Not used')]
+        self.assertErrorRegex(EasyBuildError, error_pat, ft.apply_regex_substitutions, testfile, regex_subs_part_match,
+                              on_missing_match=run.ERROR, match_all=True)
+        # First matched so OK with match_all
+        ft.apply_regex_substitutions(testfile, regex_subs_part_match,
+                                     on_missing_match=run.ERROR, match_all=False)
 
         # Warn
         with self.log_to_testlogfile():
             ft.apply_regex_substitutions(testfile, regex_subs_no_match, on_missing_match=run.WARN)
         logtxt = ft.read_file(self.logfile)
         self.assertIn('WARNING ' + error_pat, logtxt)
+        with self.log_to_testlogfile():
+            ft.apply_regex_substitutions(testfile, regex_subs_part_match, on_missing_match=run.WARN, match_all=True)
+        logtxt = ft.read_file(self.logfile)
+        self.assertIn('WARNING ' + error_pat, logtxt)
 
         # Ignore
         with self.log_to_testlogfile():
@@ -1465,6 +1476,21 @@ def test_apply_regex_substitutions(self):
         path = os.path.join(self.test_prefix, 'nosuchfile.txt')
         self.assertErrorRegex(EasyBuildError, error_pat, ft.apply_regex_substitutions, path, regex_subs)
 
+        # Replace multi-line strings
+        testtxt = "This si wrong\nBut mkae right\nLeave this!"
+        expected_testtxt = 'This is wrong.\nBut make right\nLeave this!'
+        ft.write_file(testfile, testtxt)
+        repl = ('This si( .*)\n(.*)mkae right$', 'This is wrong.\nBut make right')
+        ft.apply_regex_substitutions(testfile, [repl], backup=False, on_missing_match=ERROR, single_line=False)
+        new_testtxt = ft.read_file(testfile)
+        self.assertEqual(new_testtxt, expected_testtxt)
+        # Supports capture groups
+        ft.write_file(testfile, testtxt)
+        repl = ('This si( .*)\n(.*)mkae right$', r'This is\1.\n\2make right')
+        ft.apply_regex_substitutions(testfile, [repl], backup=False, on_missing_match=ERROR, single_line=False)
+        new_testtxt = ft.read_file(testfile)
+        self.assertEqual(new_testtxt, expected_testtxt)
+
         # make sure apply_regex_substitutions can patch files that include UTF-8 characters
         testtxt = b"foo \xe2\x80\x93 bar"  # This is an UTF-8 "-"
         ft.write_file(testfile, testtxt)
@@ -1485,34 +1511,32 @@ def test_apply_regex_substitutions(self):
 
         # also test apply_regex_substitutions with a *list* of paths
         # cfr. https://github.com/easybuilders/easybuild-framework/issues/3493
+        # and a compiled regex
         test_dir = os.path.join(self.test_prefix, 'test_dir')
         test_file1 = os.path.join(test_dir, 'one.txt')
         test_file2 = os.path.join(test_dir, 'two.txt')
         ft.write_file(test_file1, "Donald is an elephant")
         ft.write_file(test_file2, "2 + 2 = 5")
         regexs = [
-            ('Donald', 'Dumbo'),
+            (re.compile('donald', re.I), 'Dumbo'),  # Only matches if this is used as-is
             ('= 5', '= 4'),
         ]
         ft.apply_regex_substitutions([test_file1, test_file2], regexs)
 
         # also check dry run mode
         init_config(build_options={'extended_dry_run': True})
-        self.mock_stderr(True)
-        self.mock_stdout(True)
-        ft.apply_regex_substitutions([test_file1, test_file2], regexs)
-        stderr, stdout = self.get_stderr(), self.get_stdout()
-        self.mock_stderr(False)
-        self.mock_stdout(False)
+        with self.mocked_stdout_stderr():
+            ft.apply_regex_substitutions([test_file1, test_file2], regexs)
+            stderr, stdout = self.get_stderr(), self.get_stdout()
 
         self.assertFalse(stderr)
-        regex = re.compile('\n'.join([
+        regex = '\n'.join([
             r"applying regex substitutions to file\(s\): .*/test_dir/one.txt, .*/test_dir/two.txt",
-            r"  \* regex pattern 'Donald', replacement string 'Dumbo'",
+            r"  \* regex pattern 'donald', replacement string 'Dumbo'",
             r"  \* regex pattern '= 5', replacement string '= 4'",
             '',
-        ]))
-        self.assertTrue(regex.search(stdout), "Pattern '%s' should be found in: %s" % (regex.pattern, stdout))
+        ])
+        self.assertTrue(re.search(regex, stdout), "Pattern '%s' should be found in: %s" % (regex, stdout))
 
     def test_find_flexlm_license(self):
         """Test find_flexlm_license function."""