Issue #17774 has been updated by mame (Yusuke Endoh).


Thank you, I can reproduce the issue.

The issue is in the code from [onigmo](https://github.com/k-takata/Onigmo), so it would be helpful if you could report this issue to the upstream.

By a quick investigation, an optimization expands `(){4}`, and does not expand `(){5}`, which makes the difference of the behavior.
Enabling debug output suggests that the bug is caused by `USE_MONOMANIAC_CHECK_CAPTURES_IN_ENDLESS_REPEAT` option. The `(){5}` case works great by the following change that disables the option, but I'm unsure the performance impact.

```
diff --git a/regint.h b/regint.h
index 0740429688..968ea6cde8 100644
--- a/regint.h
+++ b/regint.h
@@ -71,7 +71,6 @@
 #define USE_PERL_SUBEXP_CALL
 #define USE_CAPITAL_P_NAMED_GROUP
 #define USE_BACKREF_WITH_LEVEL        /* \k<name+n>, \k<name-n> */
-#define USE_MONOMANIAC_CHECK_CAPTURES_IN_ENDLESS_REPEAT  /* /(?:()|())*\2/ */
 #define USE_NEWLINE_AT_END_OF_STRING_HAS_EMPTY_LINE     /* /\n$/ =~ "\n" */
 #define USE_WARNING_REDUNDANT_NESTED_REPEAT_OPERATOR
 /* !!! moved to regenc.h. */ /* #define USE_CRNL_AS_LINE_TERMINATOR */
```

----------------------------------------
Bug #17774: Quantified empty group causes regex to fail
https://bugs.ruby-lang.org/issues/17774#change-91317

* Author: Davidebyzero (David Ellsworth)
* Status: Open
* Priority: Normal
* ruby -v: ruby 2.7.2p137 (2020-10-01 revision 5445e04352) [x86_64-msys]
* Backport: 2.5: UNKNOWN, 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN
----------------------------------------
The regex `^((x*)(?=\2$))*x$` matches powers of 2 in unary, expressed as strings of `x` characters whose length is the number.

Adding an empty group `()` in the middle of it should have no effect on its operation, and indeed it does not. `^((x*)()(?=\2$))*x$` still matches powers of 2 just fine.  
Quantifying that empty group, `(){4}`, should still have no effect. And indeed, `^((x*)(){4}(?=\2$))*x$` still matches powers of 2. But quantify that to `(){5}`, and suddenly it fails.

The following command line should print `1`, but instead prints nothing:
```
ruby -e 'print 1 if "x"*32 =~ /^((x*)(){5}(?=\2$))*x$/'
```
However this one does print `1`:
```
ruby -e 'print 1 if "x"*32 =~ /^((x*)(){4}(?=\2$))*x$/'
```

Bug found to occur on [Try It Online](https://tio.run/): `ruby 2.5.5p157 (2019-03-15 revision 67260) [x86_64-linux]`
Bug confirmed to happen on my own machine: `ruby 2.7.2p137 (2020-10-01 revision 5445e04352) [x86_64-msys]`

Solving the challenge [Is that number a Two Bit Number?](https://codegolf.stackexchange.com/questions/211840/is-that-number-a-two-bit-number%ef%b8%8f/222792#222792) on Code Golf Stack Exchange is what led me to discover this bug.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>