Thread overview
[Issue 6888] New: std.getopt.getopt: one-letter hash option causes range violation
Nov 03, 2011
Maksim Zholudev
Nov 03, 2011
Maksim Zholudev
Nov 14, 2011
Kenji Hara
November 03, 2011
http://d.puremagic.com/issues/show_bug.cgi?id=6888

           Summary: std.getopt.getopt: one-letter hash option causes range
                    violation
           Product: D
           Version: D2
          Platform: x86_64
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Phobos
        AssignedTo: nobody@puremagic.com
        ReportedBy: maximzms@gmail.com


--- Comment #0 from Maksim Zholudev <maximzms@gmail.com> 2011-11-03 08:48:03 PDT ---
The following code causes the error:
--------------------
import std.getopt;

void main()
{
    int[string] foo;
    auto args = ["", "-t", "a=1"];
    getopt(args, "t", &foo);
}
--------------------
core.exception.RangeError@std.getopt(519): Range violation
--------------------

There is no error if bundling is turned on:
getopt(args, config.bundling, "t", &foo);

There is no error if the option contains more than one letter e.g. "--tune"

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
November 03, 2011
http://d.puremagic.com/issues/show_bug.cgi?id=6888



--- Comment #1 from Maksim Zholudev <maximzms@gmail.com> 2011-11-03 09:58:06 PDT ---
The bug is found!

In std/getopt.d:

If the option is not long and bundling is forbidden lines 587-588 are executed. If arg looks like "-t" value is empty but !value is false, i.e. in the following code all asserts are true:

test.d:
--------------------
void main()
{
    string s = "t";
    string s1 = s[1..$];
    string s2 = null;

    assert(s1 == null);
    assert(s2 == null);
    assert(!s1 == false); // Why?
    assert(!s2 == true);
}
--------------------

Hence in std/getopt.d in the line 464:
!val is false and val is empty when we come to line 519 and try to access its
contents.

I'm going to fix it right now.
But the behavior of empty slices looks like a bug of the compiler.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
November 14, 2011
http://d.puremagic.com/issues/show_bug.cgi?id=6888


Kenji Hara <k.hara.pg@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


--- Comment #2 from Kenji Hara <k.hara.pg@gmail.com> 2011-11-14 04:52:25 PST ---
https://github.com/D-Programming-Language/phobos/pull/315

https://github.com/D-Programming-Language/phobos/commit/0ee6a1afbd8686a8051af4c86df9333c22daec89

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------