FileCheck Enhancement - repeats in regular expressions

[eklepilkina]

Summary

Repeat with current number should become available by using {n}, {n,m} , {,n}, {n,}

[alexfh]

docs/CommandGuide/FileCheck.rst

485

Seems like this doesn't need a separate mention in the FileCheck documentation: it's just a feature of the regular expressions, and it should just work.

test/FileCheck/repeat-in-regex.txt

1

Will {{ˆ}}} still match a closing curly brace at the start of line? In general, there are tests that use }}} in the CHECK lines, and the change in how this sequence is parsed can affect them. Have you tried to run all LLVM and Clang tests with this change?

utils/FileCheck/FileCheck.cpp

242

How about a bit more self-documenting version?

StringRef::size_type LastBrace = PatternStr.find_first_not_of('}', EndChar);
if (LastBrace != StringRef::npos)
  EndChar = LastBrace - 1;

[eklepilkina]

docs/CommandGuide/FileCheck.rst

485

I added, because this feature doesn't work before. I think that description in documantation isn't extra. I agree that it may be replaced.

test/FileCheck/repeat-in-regex.txt

1

Yes, I did. I ran all LLVM and Clang tests with this patch.

utils/FileCheck/FileCheck.cpp

242

But if there is npos, then there won't be any shift even if last characters are }}} and End will be smaller then it should be.

docs/CommandGuide/FileCheck.rst

 
 .. code-block:: llvm
 
        DW_AT_location [DW_FORM_sec_offset]   (0x00000233)
        DW_AT_name [DW_FORM_strp]  ( .debug_str[0x000000c9] = "intd")
 
 letting us set the :program:`FileCheck` variable ``DLOC`` to the desired value 
 ``0x00000233``, extracted from the line immediately preceding "``intd``".
+
+Repeat in regular expressions
+~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+You can use quantifiers with numbers in regular expressions such as {n}, {n,m}, {n,}.
+For example:
+
+.. code-block:: llvm
+
+    // CHECK : {{matched{5}}}
+
+It’s equal to
+
+.. code-block:: llvm
+
+    // CHECK : {{matchedmatchedmatchedmatchedmatched}}

test/FileCheck/repeat-in-regex.txt

+// RUN: FileCheck -input-file %s %s
+
+text to be matchedmatchedmatchedmatchedmatched
+// CHECK: {{(matched){5}}}
+
+r1, is register is register is register is register is register
+// CHECK: [[REG:r[0-9]{1,2}]]
+// CHECK-SAME: {{(is register ){1,}}}
+
+r2r13r1
+dag
+// CHECK-DAG: {{dag{1}}}
+// CHECK-DAG: {{[[REG]]{0,2}}}

utils/FileCheck/FileCheck.cpp

if (PatternStr.startswith("{{")) {
       size_t End = PatternStr.find("}}");
       if (End == StringRef::npos) {
         SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
                         SourceMgr::DK_Error,
                         "found start of regex string with no end '}}'");
         return true;
       }
 
+      // Find end of pattern in case of quantifiers of the form {n,m}
+      size_t NextChar = End + 2;
+      while (NextChar < PatternStr.size() && PatternStr[NextChar] == '}') {
+        NextChar++;
+      }
+
+      End = NextChar - 2;
+
       // Enclose {{}} patterns in parens just like [[]] even though we're not
       // capturing the result for any purpose.  This is required in case the
       // expression contains an alternation like: CHECK:  abc{{x|z}}def.  We
       // want this to turn into: "abc(x|z)def" not "abcx|zdef".
       RegExStr += '(';
       ++CurParen;
 
       if (AddRegExToRegEx(PatternStr.substr(2, End-2), CurParen, SM))