From @cpansprout

OK, so this is the *real* bug I was after when I opened ticket #128737 (when I wasn’t thinking)​:

$ ./miniperl -e 'eval "print q \x{fffffff}Hello, \\\\whirled!\n\x{fffffff}" or die;'
Hello, \whirled!

That is expected. But add another f​:

$ ./miniperl -e 'eval "print q \x{ffffffff}Hello, \\\\whirled!\n\x{ffffffff}" or die;'
Hello, \\whirled!

And \\ in a q() string is now being preserved, instead of being changed to a single backslash.

Same results from 5.8.7 to 5.25.3.

--

Father Chrysostomos

From @cpansprout

On Tue Jul 26 00​:28​:37 2016, sprout wrote​:

OK, so this is the *real* bug I was after when I opened ticket #128737
(when I wasn’t thinking)​:

$ ./miniperl -e 'eval "print q \x{fffffff}Hello,
\\\\whirled!\n\x{fffffff}" or die;'
Hello, \whirled!

That is expected. But add another f​:

$ ./miniperl -e 'eval "print q \x{ffffffff}Hello,
\\\\whirled!\n\x{ffffffff}" or die;'
Hello, \\whirled!

And \\ in a q() string is now being preserved, instead of being
changed to a single backslash.

Same results from 5.8.7 to 5.25.3.

Another symptom of the same bug​:

\x{100000027} acts as a single quote for regexps​:

$ ./miniperl -le 'eval "print qr \x{100000027}\$^O\x{100000027}" || die'
(?^​:$^O)
$ ./miniperl -le 'eval "print qr \x{10000002}\$^O\x{10000002}" || die'
(?^​:darwin)

--

Father Chrysostomos

The RT System itself - Status changed from 'new' to 'open'

From @khwilliamson

On 07/26/2016 01​:39 AM, Father Chrysostomos via RT wrote​:

On Tue Jul 26 00​:28​:37 2016, sprout wrote​:

OK, so this is the *real* bug I was after when I opened ticket #128737
(when I wasn’t thinking)​:

$ ./miniperl -e 'eval "print q \x{fffffff}Hello,
\\\\whirled!\n\x{fffffff}" or die;'
Hello, \whirled!

That is expected. But add another f​:

$ ./miniperl -e 'eval "print q \x{ffffffff}Hello,
\\\\whirled!\n\x{ffffffff}" or die;'
Hello, \\whirled!

And \\ in a q() string is now being preserved, instead of being
changed to a single backslash.

Same results from 5.8.7 to 5.25.3.

Another symptom of the same bug​:

\x{100000027} acts as a single quote for regexps​:

$ ./miniperl -le 'eval "print qr \x{100000027}\$^O\x{100000027}" || die'
(?^​:$^O)
$ ./miniperl -le 'eval "print qr \x{10000002}\$^O\x{10000002}" || die'
(?^​:darwin)

This is because termcode in scan_str() needs to be IV and not I32.
Probably a test should be added to croak when out-of-range.

From @khwilliamson

Fixed by

commit 2aa61ac
Author​: Karl Williamson <khw@​cpan.org>
Date​: Thu Aug 4 14​:07​:53 2016 -0600

  PATCH​: [perl #128738] Large code pt as string delim
 
  The bug here is simply an I32 was used when an IV was needed.
 
  One could argue that the parser should refuse to
  accept something larger than an IV. I chose not to do that, as this is
  a deprecated usage, which generates a warning by default and will be a
  syntax error anyway in a future release.

--
Karl Williamson

@khwilliamson - Status changed from 'open' to 'pending release'

From @khwilliamson

Thank you for filing this report. You have helped make Perl better.

With the release today of Perl 5.26.0, this and 210 other issues have been
resolved.

Perl 5.26.0 may be downloaded via​:
https://metacpan.org/release/XSAWYERX/perl-5.26.0

If you find that the problem persists, feel free to reopen this ticket.

@khwilliamson - Status changed from 'pending release' to 'resolved'