Mailing-List: contact stdcxx-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: stdcxx-dev@incubator.apache.org
Received-SPF: pass (asf.osuosl.org: domain of AntonP@moscow.vdiweb.com
 designates 195.210.189.132 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: RE: test for lib.string.insert
Date: Fri, 31 Mar 2006 20:19:13 +0400
Message-ID: <4D6A8407B7AC6F4D95B0E55C4E7C4C6203F600FC@exmsk.moscow.vdiweb.com>
Thread-Topic: test for lib.string.insert
Thread-Index: AcZUi7jdrNTgP7MrQX+6EU7/iFyUvAATYawA
From: "Anton Pevtsov" <AntonP@moscow.vdiweb.com>
To: <stdcxx-dev@incubator.apache.org>

Martin Sebor wrote:
> I see a number of failed assertions when I compile the test with gcc
4.1.0 and run it on Solaris (also attached). Are those expected? I.e.,
is there actually a problem with the library? If so (i.e., if > you can
reproduce the problem), could you put together a small test case and
open an issue in Jira?

These assertions are not expected, but their cause is the problem with
rw_widen function for char's. But I suspect another problem with insert
(s.begin(), s.begin (), s.end ()) - I prepare a small test when I could
reproduce the problem separately from the string.insert test. And this
problem appears on Linux only.

Martin Sebor wrote:
> Could you please enhance the 0.char.cpp test to exercise this and
commit both the enhanced test as well as your patch?
Hm, this issue is not a bug and the 0.char.cpp test seems to be correct,
but I'll update it to exercise rw_widen on arrays without '\0' at the
end.

Thanks,
Anton Pevtsov

-----Original Message-----
From: Martin Sebor [mailto:sebor@roguewave.com]=20
Sent: Friday, March 31, 2006 06:29
To: stdcxx-dev@incubator.apache.org
Subject: Re: test for lib.string.insert


Anton Pevtsov wrote:
> The attached file contains new test version updated according your=20
> notes.
>=20
> I used different array here, but may be it will be useful to create=20
> another one using bitmaps and compare these versions.

I made a simplifying change to get_insert_format (I used the %{+}
directive to append directly to the buffer). See the attached diff. I
also removed all trailing blanks (not shown in the diff). With these
changes please go ahead and commit the test.

I see a number of failed assertions when I compile the test with gcc
4.1.0 and run it on Solaris (also attached). Are those expected? I.e.,
is there actually a problem with the library? If so (i.e., if you can
reproduce the problem), could you put together a small test case and
open an issue in Jira?

>=20
> I found an issue in the rw_widen function implementation for chars:
[...]
> I corrected this by adding '\0' symbol to the end of the resulting=20
> array (char.cpp, line 248)
>=20
> ...
>            // copy src into dst
>            memcpy (dst, src, len);
>            dst [len] =3D '\0';
> ...
>=20
> What do you think about it?

Could you please enhance the 0.char.cpp test to exercise this and commit
both the enhanced test as well as your patch?

>=20
> Also I split rw_assert's formatting parameters into three parts using
> rw_asnprintf function.
>=20
> Current test version doesn't contain exception safety tests. I am=20
> working on them.

Sounds good.

While I'm gone, please continue to work on the tests. You might as well
commit your work (tests only) so you don't accidentally lose anything.
I'll review it as I have time or when I get back.

Thanks!
Martin