August 2005 - Boost - lists.boost.org

[1.33 RC2/serialization] New failures on Intel 8.1
by Stefan Slapeta 11 Aug '05

11 Aug '05

Hi, the latest changes seem to cause new failures on Intel 8.1 ("internal failure...") that are once again related to the buggy debug info generation in this compiler version. After investigating, I'm afraid I can't offer anything but to mark them as expected :-( Stefan

2 1

Re: [boost] alignment problem in proposed any alternative
by felipe 11 Aug '05

11 Aug '05

> Date: Thu, 11 Aug 2005 09:00:20 -0400 > From: christopher diggins <cdiggins(a)videotron.ca> > Subject: Re: [boost] alignment problem in proposed any alternative > To: boost(a)lists.boost.org > Message-ID: <004c01c59e74$a17a7800$2b792518@heronnest> > Content-Type: text/plain; format=flowed; charset=iso-8859-1; > reply-type=original > ----- Original Message ----- > From: "felipe" <pbr(a)taric.es> > To: <boost(a)lists.boost.org> > Sent: Thursday, August 11, 2005 4:09 AM > Subject: Re: [boost] alignment problem in proposed any alternative > >> Nope, this is totally wrong! chars and arrays of type char are >> generally >> the least aligned objects. You can demonstrate that this doesn't work >> with something like this: >> >> void blah() >> { >> char c[/* some number */]; >> any<8> a((double)1.0); >> std::cout << &a << std::endl; >> } > That code is completely misleading, yhe question is not what the address > of > a is! > Well, let me clarify: /*############ BEGIN ########*/ #include <iostream> #include <boost/type_traits.hpp> template<int buffer_size> struct any { char c; // this offsets buffer by one-byte.. because for the // compiler, chars only need byte alignment char buffer[buffer_size]; // ... template<typename T> any(const T& x) { if (boost::is_pod<T>::value && (sizeof(T) <= buffer_size)) { memcpy(buffer, &x, sizeof(T)); } else { // new and stuff } } }; int main(int argc, char **argv) { any<8> blah((double)1.0); std::cout << (void *)&blah.buffer << std::endl; return 0; } /*##### END ########*/ Output of the program on Pentium/Win2k/Visual C++ 6.0: 0x22ef01 This is obviously NOT a good place for a double, unless you are going to memcpy everything, which would be a needlessly inefficient design. Right, the Pentium lets me slide (with some extra instructions to keep away idle cycles), but don't try it on an Alpha! Have you not read the articles by Alexandrescu in CUJ? These are pretty old and yet they explore all of this stuff. What you should do is make the buffer a union of a minimal size object (e.g. your char array) and a minimum alignment object (e.g. a double in the example above). The compiler then adjusts alignment for the latter, while guaranteeing space for the former. Another thing, your proposal shies away from objects with constructors, etc. Why? You can use the placement new operator and get a much cleaner implementation. You know, something like this: template<typename T> any(const T& x) { .. new(&buffer[0]) T(x); // instead of memcpy hope this helps -- Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

1 0

Self Contained Transportable Class :
by Mottengte, Rajeev Kumar 11 Aug '05

11 Aug '05

Hi I am trying to create a self contained transportable class which can be migrated to any machine by calling a member function in it. I am having the socket stream implementation which derives the transportable boost class. In the serialize function of the boost class, I am trying to get the string stream and trying to call network marshal function of the base class. How can I get the stream ( I am using a string stream ) from the archive class inside the serialize function. Rgds Rajeev

1 0

Re: [boost] alignment problem in proposed any alternative
by felipe 11 Aug '05

11 Aug '05

Nope, this is totally wrong! chars and arrays of type char are generally the least aligned objects. You can demonstrate that this doesn't work with something like this: void blah() { char c[/* some number */]; any<8> a((double)1.0); std::cout << &a << std::endl; } Tweak the size of c[] until you get &a % 8 != 0. Some compilers optimize stack layout, over-aligning data, but this is true only for SOME compilers on SOME processors with SOME optimization settings. (You can also experiment by defining an additional char variable before the buffer in your any. Where does &a.buffer lie?) I think you are misreading 3.9.2: it is talking about pointers, namely that void * and char * have the same storage and alignment requirements. On Wed, 10 Aug 2005 20:27:16 +0200, <boost-request(a)lists.boost.org> wrote: > From: "christopher diggins" <cdiggins(a)videotron.ca> > To: "Boost mailing list" <boost(a)lists.boost.org> > Sent: Wednesday, August 10, 2005 1:30 AM > Subject: [boost] alignment problem in proposed any alternative >> The question I have is whether in practice though are there are any >> compilers which have stricter alignment requirements on small types (<= >> sizeof(void*)) than a void pointer. In other words, is it possible with >> any >> known compiler to actually throw the runtime error ? > Now a moot point, see below. > >> Unfortunately there is another problem, I am not an expert in >> interepreting >> the standard, but AFAICT placement new is guaranteed to return a pointer >> to >> properly aligned memory (3.7.3.1 / 2) which I leverage to check that >> alignment occured. Apparently not every STL implementation does in fact >> do >> that (I believe at least dinkumware is guilty of that). > Apparently I am wrong about this! 18.4.1.3 provides an exception. (I > wonder > why I missed the exception 15 chapters later?) > Anyway it turns out we can guarantee alignment by simply using a char > buffer. According to 3.9/2 the following automatically has no alignment > problems: > template<int buffer_size> > struct any { > char buffer[buffer_size]; > ... > template<typename T> > any(const T& x) { > if (boost::is_pod<T>::value && (sizeof(T) <= buffer_size)) { > memcpy(buffer, &T, N); > } > else { > // new and stuff > } > } > } > Christopher Diggins -- Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

3 2

[1.33.0] Release candidate #2 is available
by Douglas Gregor 11 Aug '05

11 Aug '05

Boost 1.33.0 release candidate #2 is now available at the usual place: http://www.osl.iu.edu/~dgregor/boost-1.33.0/ Please build it, test it, and report back to the mailing list. If no major problems are found before August 9th, this release candidate will become Boost 1.33.0. Doug Gregor 1.33.0 Release Manager

12 19

Compilation error with regex only with -O3 on gcc
by robert kraus 11 Aug '05

11 Aug '05

Hi, when I compile with: -O0 to -O2 my programs compile and work fine, but when I switch to -O3 I get compilation errors (shortened, see below). Is this my fault (maybe I shouldnt use optimizations? thank you very much for your answers ido fedora core 4,gcc version: Target: i386-redhat-linux Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-libgcj-multifile --enable-languages=c,c++,objc,java,f95,ada --enable-java-awt=gtk --with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --host=i386-redhat-linux Thread model: posix gcc version 4.0.1 20050727 (Red Hat 4.0.1-5) boost: boost-devel-1.32.0-6 Errors: Building file: ../src/Parser.cpp Invoking: GCC C++ Compiler g++ -DDEBUG -O3 -g3 -pedantic -Wall -c -fmessage-length=0 -osrc/Parser.o ../src/Parser.cpp /usr/include/boost/regex/v4/regex_compile.hpp: In member function ‘boost::re_detail::re_syntax_base* boost::reg_expression<charT, traits, Allocator>::compile_set_simple(boost::re_detail::re_syntax_base*, long unsigned int, bool) [with charT = char, traits = boost::regex_traits<char>, Allocator = std::allocator<char>]’: /usr/include/boost/regex/v4/regex_compile.hpp:1635: instantiated from ‘unsigned int boost::reg_expression<charT, traits, Allocator>::set_expression(const charT*, const charT*, boost::regex_constants::syntax_option_type) [with charT = char, traits = boost::regex_traits<char>, Allocator = std::allocator<char>]’ /usr/include/boost/regex/v4/basic_regex.hpp:238: instantiated from ‘unsigned int boost::reg_expression<charT, traits, Allocator>::set_expression(const charT*, boost::regex_constants::syntax_option_type) [with charT = char, traits = boost::regex_traits<char>, Allocator = std::allocator<char>]’ /usr/include/boost/regex/v4/regex_compile.hpp:110: instantiated from ‘boost::reg_expression<charT, traits, Allocator>::reg_expression(const charT*, boost::regex_constants::syntax_option_type, const Allocator&) [with charT = char, traits = boost::regex_traits<char>, Allocator = std::allocator<char>]’ /usr/include/boost/regex/v4/basic_regex.hpp:340: instantiated from ‘boost::basic_regex<charT, traits, Allocator>::basic_regex(const charT*, typename boost::reg_expression<charT, traits, Allocator>::flag_type, const Allocator&) [with charT = char, traits = boost::regex_traits<char>, Allocator = std::allocator<char>]’ ../src/Parser.cpp:16: instantiated from here __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

3 2

Re: [boost] [format] Inconsistency with printf
by martin.ecker＠tab.at 11 Aug '05

11 Aug '05

Sorry about the previous, empty e-mail. We were having mail server problems at my company. Below is the content of my original mail: Hello, The following sample program: #include <cstdio> #include <iostream> #include <limits> #include <boost/format.hpp> int main(int argc, char* argv[]) { std::cout << boost::format("%f") % std::numeric_limits<float >::infinity() << std::endl; std::printf("%f", std::numeric_limits<float>::infinity()); return 0; } produces 1.00000000000000000000000000....00000000000000000.#INF00 with Boost.Format, but 1.#INF00 with printf. Is this a bug or shouldn't I expect the same behavior from Boost.Format for special floating point values such as NaN and infinity? Note that using %g instead of %f, that is: std::cout << boost::format("%g") % std::numeric_limits<float>::infinity() << std::endl; prints the same thing as printf, i.e. 1.#INF00 This is using Boost 1.32.0 with MSVC 7.1 on Windows. Best Regards, Martin TAB Austria Haiderstraße 40 4052 Ansfelden Austria Phone: +43 7229 78040-218 Fax: +43 7229 78040-209 E-mail: martin.ecker(a)tab.at http://www.tab.at

1 0

Re: [boost] surveying async network io libraries
by Matthew Nourse 11 Aug '05

11 Aug '05

>On 8/10/05, Jose <jmalv04_at_[hidden]> wrote: > >The asio library really looks quite nice. It appears simple to use, >is 100% header-based, doesn't require the use of threads, and uses the >most scalable event demultiplexor supplied by the OS (epoll on Linux >in the 0.3x version and IOCP on Windows). Its a very nice little >library! Thanks for linking to it. I think I'm in love :-) I've been writing asio apps for a closed-source software company for the last 16 months. With asio, I've written: * a custom HTTP proxy that has been running in production non-stop since October last year * an HTTP<->XMPP gateway * a simple asio-based HTTP load test tool (so I can simulate 20,000 simultaneous HTTP "clients" from a single process) * a publish/subscribe server with a simple binary protocol- am still trying to performance tune other code so I can find the limits of this, so far 500 messages/sec doesn't cause any detectable CPU usage on a 3GHz P4. * A second custom HTTP proxy- this one does some authentication as well. The first application runs on Windows and the rest run on Debian Linux. They all share HTTP/networking code where applicable. All applications use boost::bind and boost::shared_ptr extensively and the publish/subscribe server uses boost::regex to figure out how to route requests. I am addicted to asio. I can't stop using it ;-). Matt

1 0

Interest for a perl6-like-junctions-class?
by FlSt＠gmx.de 11 Aug '05

11 Aug '05

Hi! The new Perl 6 has a nice feature called junctions. See http://www.perl.com/pub/a/2003/07/29/exegesis6.html?page=4 and http://www.metaperl.com/talks/p6-junctions/ for more information. I think this is a interesting way to save the time for writing loops in some situations. I've tried to implement this for C++ and uploaded a version with two examples into the Boost Sandbox File Vault (flos_super_position.zip). This version works for my purposes, but if you interested i would rework it and submit it to the boost library. Sincerely, Florian S.

13 95

shared_ptr and threads
by Timothy Ritchey 11 Aug '05

11 Aug '05

We are using shared_ptr extensively in our application, and seem to be having some threading issues. In particular, we have an Envelope class that is used to hold messages received over the network. We use Spirit to parse the message string and construct the Envelope object, then we pass is on to the EnvelopeController, which holds a thread pool to handle all of the Envelopes. These Envelopes are passed around as shared_ptr<Envelope> between these different components. In essense, we have one thread that is receiving the message from the network stack, parsing it, and passing it to the thread_pool, which is then using one (or more) additional threads to actually deal with the contents of the envelope. On a Uni-processor box, everything works great. All of our envelopes are being destroyed properly. On a multi-processor machine, none of the envelopes are getting destroyed. We are in a situation where two shared_ptrs to the same underlying object may go out of scope and need to decrement the ref count at the same time. Is the in the realm of "undefined behavior?" Am I using the shared_ptr incorrectly? Thanks for any advice. Thread 1 --------- { shared_ptr<Envelope> envelope(new Envelope(...)); ... [pass envelope to Thread 2] ... [envelope goes out of scope] } Thread 2 -------- void process(shared_ptr<Envelope> envelope) { [do work on envelope] [envelope goes out of scope] }

2 2