Mailing-List: contact stdcxx-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: stdcxx-dev@incubator.apache.org
Received-SPF: pass (nike.apache.org: domain of Farid_Zaripov@epam.com
 designates 217.21.63.3 as permitted sender)
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: 22.locale.money.get.cpp asserts
Date: Tue, 25 Sep 2007 21:24:44 +0300
Message-ID: <7BDB2168BEAEF14C98F1901FD2DE64380100922A@epmsa009.minsk.epam.com>
Thread-Topic: 22.locale.money.get.cpp asserts
Thread-Index: Acf/oViuwt/uePZ6SxygRiB0AD7JiQ==
From: "Farid Zaripov" <Farid_Zaripov@epam.com>
To: <stdcxx-dev@incubator.apache.org>

  I have checked the re-asserting lines of the test and compared the
results with
DinkumwareSTL and STLport.

  The line 752:

    // all spaces extracted since currency symbol (last) is mandatory
    // verify that the facet doesn't extract too many optional spaces
    // leaving none for the final required currency symbol
    TEST (T,  103.0, "103 ", 4, showbase, eofbit, 0, "\3\4\0\2", " ");

  The line 798:

    const charT minus_space[][3] =3D { { '-', ' ', '\0' }, { '\0' } };
    PunctData<charT>::negative_sign_ [intl] =3D minus_space [0];
    PunctData<charT>::negative_sign_ [locl] =3D minus_space [1];

    // { sign, value, none, symbol }
    // negative_sign is "- ", (note the single trailing space)
    // verify that the extraction suceeds, i.e., that the
money_base::none
    // specifier that's last in the pattern doesn't confuse the facet
into
    // extracting all the optional whitespace, leaving none to complete
    // the negative_sign
    TEST (T, -109.1, "-109  ", 6, 0, eofbit, 0, "\3\4\0\2", "");

  The problem is in that all spaces are eaten when 'none' pattern
element was parsed,
but the ending currency character in first test and trailing
negative_sign character
in second test are requires for presence of the space character.

  The both test lines above are failed on DinkumwareSTL and STLport too.
The questions:
- may the negative_sign have the trailing part with only space
characters?
- may the currency symbol contain only spaces?


  The another lines with rw_asserts() (851 - 853):

    PunctData<charT>::positive_sign_ [intl] =3D 0;
    PunctData<charT>::positive_sign_ [locl] =3D 0;
    PunctData<charT>::negative_sign_ [intl] =3D 0;
    PunctData<charT>::negative_sign_ [locl] =3D 0;

    [...]

    TEST (T,  117.0, "117$",   4, showbase, eofbit,  0, "\4\0\2\3",
"$");
    TEST (T,  118.0, "118$",   4, showbase, eofbit,  0, "\4\0\2\3",
"$");
    TEST (T,  119.0, "119 $",  5, showbase, eofbit,  0, "\4\1\2\3",
"$");

  Here the sign is the last position in pattern, and positive sign and
negative sign both are empty.
I've prepared the patch, fixing the problem:


Index: _money_get.cc
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- _money_get.cc	(revision 579104)
+++ _money_get.cc	(working copy)
@@ -165,7 +165,10 @@
         case /* '\3' */ money_base::sign: {
=20
             if (__it =3D=3D __end) {
-                __ebits |=3D _RW::__rw_failbit;
+                if (__ps.size () && __ns.size ())
+                    __ebits |=3D _RW::__rw_failbit;
+                else
+                    __sign =3D __ps.empty () - 1;
                 break;
             }
=20

Farid.