[pcre-dev] [Bug 1801] New: Test #18 fails

Top Page
Delete this message
Author: admin
Date:  
To: pcre-dev
Subject: [pcre-dev] [Bug 1801] New: Test #18 fails
https://bugs.exim.org/show_bug.cgi?id=1801

            Bug ID: 1801
           Summary: Test #18 fails
           Product: PCRE
           Version: 10.21 (PCRE2)
          Hardware: x86
                OS: Windows
            Status: NEW
          Severity: bug
          Priority: medium
         Component: Code
          Assignee: ph10@???
          Reporter: Roy.Ivy.III@???
                CC: pcre-dev@???


I recently updated and compiled the most recent commit of PCRE2
(svn://vcs.exim.org/pcre2/code/trunk@493).

When compiled under Windows/MinGW, test #18 fails (only for the 8-bit library;
16-bit and 32-bit are purposefully skipped).

The relevant mismatch in "testout8\testoutput18" is:

```
...

/abcd/substitute_extended
** Ignored with POSIX interface: substitute_extended

/\[A]{1000000}**/expand,regerror_buffsize=31
Failed: POSIX code 4: ? * + invalid at offset 1000001DEADBEEFAAAAAAAAAAA ...
(~1 million repeats)
** regerror() message truncated

/\[A]{1000000}**/expand,regerror_buffsize=32
Failed: POSIX code 4: ? * + invalid at offset 1000001

# End of testdata/testinput18
```

vs the expected (snipped):

```
...

/abcd/substitute_extended
** Ignored with POSIX interface: substitute_extended

/\[A]{1000000}**/expand,regerror_buffsize=31
Failed: POSIX code 4: ? * + invalid at offset 100000
** regerror() message truncated

/\[A]{1000000}**/expand,regerror_buffsize=32
Failed: POSIX code 4: ? * + invalid at offset 1000001

# End of testdata/testinput18
```

It looks like it might be a buffer overflow error.

I looked back into the SVN history and the bug is introduced in commit
svn://vcs.exim.org/pcre2/code/trunk@399, apparently present at the introduction
of the test. Unfortunately, I don't know the code well enough to narrow it down
further.

All other tests (8-bit, 16-bit, and 32-bit libraries, using both 32-bit and
64-bit compilers) are passing, except for the locale-specific tests (which I
expect to fail).

--
You are receiving this mail because:
You are on the CC list for the bug.