Boost logo

Boost :

From: Boris Gubenko (Boris.Gubenko_at_[hidden])
Date: 2007-10-16 20:18:02


Eric Niebler wrote:
> Hi, Boris. Is there any news about this issue? Can you suggest a
> work-around or should I mark these failures expected?

Hi, Eric. No progress, sorry. Did not have any chance to look at
this.

I think that marking these failures expected with a comment like
(stolen from John Maddock) "These test failures are reported to be
under investigation at HP's compiler lab" is the right thing to do.

Please be assured that this issue is not forgotten and when I have
anything to report, I'll do so.

Thanks,
  Boris

----- Original Message -----
From: "Eric Niebler" <eric_at_[hidden]>
To: <boost_at_[hidden]>
Cc: "Boris Gubenko" <Boris.Gubenko_at_[hidden]>
Sent: Tuesday, October 16, 2007 7:49 PM
Subject: Re: [boost] [xpressive] HP-UX_ia64_aCC problem ... care to help?

>
> Boris Gubenko wrote:
>> test_actions crashes when std::string is trying to dereference _C_data
>> member which happens to be zero. It happens inside regex_match().
>> Attached slightly reduced reproducer.
>>
>> test_symbols.cpp has a similar call to regex_match() and I'm sure it
>> crashes for the same reason.
>>
>> Just rearranging the code in RW library <string> header makes the
>> crash disappear which may be indicative of a code generation problem.
>> And everything is fine when compiling with g++ 4.2.1.
>>
>> I suggest, unless you have a reason to believe that it may be some
>> code in xpressive that is causing the problem, we treat these test
>> failures as aC++ code generation problem at this point. We'll
>> investigate it and report back.
>>
>> #include <boost/xpressive/xpressive_static.hpp>
>> #include <boost/xpressive/regex_actions.hpp>
>> extern "C" int puts(const char*);
>>
>> namespace xp = boost::xpressive;
>>
>> int main()
>> {
>> using namespace boost::xpressive;
>>
>> std::string result;
>> std::string str("foo bar");
>> sregex rx = (+_w)[ xp::ref(result) += _ ] >> *(' ' >> (+_w)[
>> xp::ref(result)
>> += ',' + _ ]);
>> puts("calling regex_match");
>> regex_match(str, rx);
>> puts("exiting");
>> }
>>
>
>
>
> Hi, Boris. Is there any news about this issue? Can you suggest a
> work-around or should I mark these failures expected?
>
>
>
> --
> Eric Niebler
> Boost Consulting
> www.boost-consulting.com
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk