keygeneration on ios using clang(libc++) never terminates - Crypto++ Users

The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Message from discussion keygeneration on ios using clang(libc++) never terminates
Received-SPF: pass (google.com: domain of fraser.hutchi...@gmail.com designates 74.125.82.46 as permitted sender) client-ip=74.125.82.46;
Message-ID: <508EE9E6.6080...@googlemail.com>
Date: Mon, 29 Oct 2012 20:41:10 +0000
From: Fraser Hutchison <fraser.hutchi...@gmail.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:16.0) Gecko/20121010 Thunderbird/16.0.1
MIME-Version: 1.0
To: Marshall Clow <mclow.li...@gmail.com>
CC: "cryptopp-users@googlegroups.com" <cryptopp-users@googlegroups.com>
Subject: Re: keygeneration on ios using clang(libc++) never terminates
References: <34614981.p...@talk.nabble.com> <C7A46DDA-27BB-4300-A9CE-F1530E544...@gmail.com>
In-Reply-To: <C7A46DDA-27BB-4300-A9CE-F1530E544...@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


On 29/10/2012 14:39, Marshall Clow wrote:
> On Oct 29, 2012, at 6:08 AM, thwang <nab...@wang-nielsen.dk> wrote:
>
>> Hi all
>>
>> I have recently change compiler, from gcc(libstdc++) to clang(libc++) on
>> iOS. After the change the generation of key, using the following snib, never
>> terminates.
>> ...
>> CryptoPP::RSAES< CryptoPP::OAEP <CryptoPP::SHA1> >::PrivateKey
>> privateSessionKey;
>> CryptoPP::AutoSeededX917RNG<CryptoPP::AES> rng;
>> privateSessionKey.GenerateRandomWithKeySize(rng, 2048);
>> ...
>>
>> I haunted it dow to be related to the use of vector<bool> in
>> nbtheory.cpp(line 310) which seems to give unexpected values when using the
>> clang(and libc++).
>> ..
>> 308 bool PrimeSieve::NextCandidate(Integer &c)
>> 309 {
>> 310      bool safe = SafeConvert(std::find(m_sieve.begin()+m_next,
>> m_sieve.end(), false) - m_sieve.begin(), m_next);
>> 311        assert(safe);
>> 312      if (m_next == m_sieve.size())
>> ..
>>
>> I isolated the issue to the following example:
>> std::vector<bool> my_vector;
>> my_vector.push_back(true);
>> my_vector.push_back(false);
>> my_vector.push_back(true);
>> my_vector.push_back(true);
>> my_vector.push_back(true);
>> my_vector.push_back(false);
>> my_vector.push_back(true);
>> long idx = std::find(my_vector.begin()+2, my_vector.end(), false) -
>> my_vector.begin();
>> std::cout << "idx:" << idx << std::endl;
>>
>> which returns
>> idx:0
>> when using clang (and libc++) and
>> idx:5
>> when using gcc (and libstdc++) as expected.
>>
>> Can anybody confirm this issue?
> Nice catch! - thanks for the simple test case.
>
> I can confirm that this issue is present in the shipping libc++ (tested with Xcode 4.5.1)
> I can also confirm that this issue has been fixed in the libc++ source repository at:
> 	https://llvm.org/svn/llvm-project/libcxx/trunk
>
> A short term fix would be to replace just the file "__bit_reference" in your Xcode installation (after backing up, of course) with the one from the libc++ source repo.
> On my system, that's in:
> 	/Applications/Xcode.app/Contents/Developer/Toolchain/XcodeDefault.xctoolchain/usr/lib/c++/v1/__bit_reference
>
> -- Marshall
>
> Marshall Clow     Idio Software   <mailto:mclow.li...@gmail.com>
>
> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait).
>          -- Yu Suzuki
>

Hi there,

I reported this a while back: 
http://sourceforge.net/apps/trac/cryptopp/ticket/17

It looks like there's not much activity in the bug tracker - should I 
have reported this elsewhere?

Thanks,
Fraser.
Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
Account Options
