- Boost - lists.boost.org

[multi_index] benchmarking
by Maxim Yegorushkin 28 May '06

28 May '06

As a part of a project I needed to compare performances of boost::multi_index hashed_unique/ordered_unique indexes with those of std::tr1::unordered_set, __gnu_cxx::hash_set, std::set. multi_index performance documentation page does not provide this kind of information, so I had to write my own test. I thought it might be a good idea to add the results to the documentation page. Source code for the test is provided. Here is a sample run: [max@localhost cont]$ sudo bin/test -q -n 10000 -x 10 10000 elements; 10 iterations insert test (lower is better) ext_hash [ 44.43%] OOOOOOOOOOOOOOOOO 29714 usec std_hash [ 50.90%] OOOOOOOOOOOOOOOOOOOO 34042 usec mi_hash [ 69.61%] OOOOOOOOOOOOOOOOOOOOOOOOOOO 46561 usec mi_set [ 97.17%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 64990 usec std_set [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 66885 usec find hit test (lower is better) ext_hash [ 22.54%] OOOOOOOOO 7266 usec std_hash [ 23.87%] OOOOOOOOO 7694 usec mi_hash [ 29.09%] OOOOOOOOOOO 9378 usec std_set [ 87.99%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 28364 usec mi_set [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 32235 usec find miss test (lower is better) std_set [ 85.09%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 8027 usec ext_hash [ 93.33%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 8805 usec mi_set [ 93.69%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 8839 usec std_hash [ 97.62%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 9209 usec mi_hash [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 9434 usec iterate test (lower is better) std_hash [ 43.66%] OOOOOOOOOOOOOOOOO 3129 usec mi_hash [ 55.60%] OOOOOOOOOOOOOOOOOOOOOO 3984 usec mi_set [ 57.26%] OOOOOOOOOOOOOOOOOOOOOO 4103 usec std_set [ 93.54%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 6703 usec ext_hash [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 7166 usec erase test (lower is better) std_hash [ 27.31%] OOOOOOOOOO 23914 usec ext_hash [ 28.68%] OOOOOOOOOOO 25117 usec mi_hash [ 38.02%] OOOOOOOOOOOOOOO 33299 usec mi_set [ 92.56%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 81057 usec std_set [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 87573 usec memory test (lower is better) std_hash [ 60.54%] OOOOOOOOOOOOOOOOOOOOOOOO 121096 bytes ext_hash [ 64.57%] OOOOOOOOOOOOOOOOOOOOOOOOO 129156 bytes mi_hash [ 64.58%] OOOOOOOOOOOOOOOOOOOOOOOOO 129168 bytes std_set [ 99.99%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 200000 bytes mi_set [100.00%] OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 200020 bytes dir := $(CURDIR) bin_dir := $(dir)/bin obj_dir := $(dir)/obj src_dir := $(CURDIR) all : test_obj := ${addprefix $(obj_dir)/,test.o} $(bin_dir)/test : $(test_obj) bin_targets += $(bin_dir)/test CPPFLAGS := CXXCOMMON := -fmessage-length=0 -Wall -Wextra -Wno-missing-field-initializers ifndef DEBUG CXXFLAGS := $(CXXCOMMON) -O3 -march=pentium-m -fearly-inlining else CXXFLAGS := $(CXXCOMMON) -ggdb endif LDDIRS := LDFLAGS := # allow -l in prerequisites VPATH := $(LDDIRS) all : $(bin_dir) $(obj_dir) $(bin_targets) $(bin_dir)/% : $(CXX) $(LDFLAGS) -o $@ $+ $(obj_dir)/%.o : %.cc $(CXX) -c $< $(CPPFLAGS) $(CXXFLAGS) -MMD -MF $(@:.o=.d) -o $@ $(bin_dir) $(obj_dir) $(lib_dir) : mkdir -p $@ # suppress built-in rules for the following targets Makefile : ; make.% : ; $(obj_dir)/%.d : ; %.hpp %.h %.cc %.cpp : ; %:: s.% ; .PHONY: all clean clean : rm -rf $(obj_dir) $(bin_dir) objs := $($(bin_targets:$(bin_dir)/%=%_obj)) deps := $(objs:.o=.d) -include $(deps)

1 0

Proposal for aiding trust issues through object design.
by Christopher Granade 27 May '06

27 May '06

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 It seems to be a recurring problem that, in designing secure server-side software, it is difficult to distinguish between variables containing trusted and untrusted input. I would propose the addition of a library to Boost that attempts to alleviate these kinds of issues by providing a template for "trusted types," as well as methods that can be marked as requiring trusted input. For instance: boost::trust::trusted<std::string> ts; cin >> ts; cout << ts.trusted(); // returns false escape_string(ts); cout << ts.trusted(); // returns true foo(ts); where foo() would have a prototype looking like: void foo(boost::trust::trusted<std::string, MUST_TRUST> str); I have not written any actual code to back up this idea, but rather wanted to propose the idea to the list first. The idea for how this would work hinges on trusted<T> offering two casting operators: one to const T and one to T. Any use of the non-const cast would result in the trust flag being set to untrusted. If anyone is interested, please let me know. Thanks. - --Christopher Granade Note: Sorry if this post is received twice. The first didn't seem to go through. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFEduB70dXuuZr00J4RAgpCAKCO1z4OByvzRUc5oiVAKxfSH/2WOQCgpQM/ IUmUZl7us1on2N1kIPW8BRk= =CX5j -----END PGP SIGNATURE-----

2 3

CVS access problem
by Robert Ramey 27 May '06

27 May '06

I'm getting the following when I try to update from the CVS system. `@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ ' from cvs server `@ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @ ' from cvs server `@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ ' from cvs server `IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! ' from cvs server `Someone could be eavesdropping on you right now (man-in-the-middle attack)! ' from cvs server `It is also possible that the RSA host key has just been changed. ' from cvs server `The fingerprint for the RSA key sent by the remote host is' from cvs server `13:f1:65:c3:6c:b7:7e:a5:f0:f3:f5:19:f4:42:9c:4a. ' from cvs server `Please contact your system administrator. ' from cvs server `Add correct host key in /home/Robert Ramey/.ssh/known_hosts to get rid of this message. ' from cvs server `Offending key in /home/Robert Ramey/.ssh/known_hosts:3 ' from cvs server `RSA host key for boost.cvs.sourceforge.net has changed and you have requested strict checking. ' from cvs server `Host key verification failed. ' from cvs server cvs [update aborted]: end of file from server (consult above messages if any) ' from cvs serverng: unrecognized response `Host key verification failed. cvs [update aborted]: end of file from server (consult above messages if any) I havn't used CVS for a couple of weeks - I did have to change my $CVSROOT string some time ago. But this is something new. Can anyone point me in the right direction to fix this? Robert Ramey

4 3

CVS problems
by Robert Ramey 27 May '06

27 May '06

I'm getting the following when I try to update from the CVS system. `@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ ' from cvs server `@ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @ ' from cvs server `@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ ' from cvs server `IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! ' from cvs server `Someone could be eavesdropping on you right now (man-in-the-middle attack)! ' from cvs server `It is also possible that the RSA host key has just been changed. ' from cvs server `The fingerprint for the RSA key sent by the remote host is' from cvs server `13:f1:65:c3:6c:b7:7e:a5:f0:f3:f5:19:f4:42:9c:4a. ' from cvs server `Please contact your system administrator. ' from cvs server `Add correct host key in /home/Robert Ramey/.ssh/known_hosts to get rid of this message. ' from cvs server `Offending key in /home/Robert Ramey/.ssh/known_hosts:3 ' from cvs server `RSA host key for boost.cvs.sourceforge.net has changed and you have requested strict checking. ' from cvs server `Host key verification failed. ' from cvs server cvs [update aborted]: end of file from server (consult above messages if any) ' from cvs serverng: unrecognized response `Host key verification failed. cvs [update aborted]: end of file from server (consult above messages if any) I havn't used CVS for a couple of weeks - I did have to change my $CVSROOT string some time ago. But this is something new. Can anyone point me in the right direction to fix this? Robert Ramey

2 2

integral_promotion.hpp & old EDG problem
by Ralf W. Grosse-Kunstleve 27 May '06

27 May '06

The new integral_promotion.hpp doesn't compile with the old EDG238-based MIPSpro 7.3.1 compiler (SGI, IRIX). This breaks our Boost.Python build. However, with the simple patch below the build is OK again. I'd love to maintain the MIPSpro Boost.Python build, but it is not worth serious work. Therefore the patch is simply #ifdef'ing out the two promotion header files for this specific platform. Are there any objections against this approach? Thanks! Ralf Index: type_traits.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/type_traits.hpp,v retrieving revision 1.16 diff -u -r1.16 type_traits.hpp --- type_traits.hpp 22 May 2006 21:02:59 -0000 1.16 +++ type_traits.hpp 26 May 2006 22:45:51 -0000 @@ -71,8 +71,10 @@ #include "boost/type_traits/function_traits.hpp" #include "boost/type_traits/aligned_storage.hpp" #include "boost/type_traits/floating_point_promotion.hpp" +#if !(defined(__sgi) && defined(__EDG_VERSION__) && (__EDG_VERSION__ == 238)) #include "boost/type_traits/integral_promotion.hpp" #include "boost/type_traits/promote.hpp" +#endif #include "boost/type_traits/ice.hpp" For the records, this is the error: cc-3313 CC: ERROR File = boost/boost/type_traits/integral_promotion.hpp, Line = 158 The expression must have arithmetic or enum type. BOOST_STATIC_CONSTANT(int, value = sizeof(promoted_index_tester(+testee)) ); ^ I've reduced it to: template<int N> struct sized_type_for_promotion { typedef char (&type)[N]; }; sized_type_for_promotion<1> ::type promoted_index_tester(int); sized_type_for_promotion<2> ::type promoted_index_tester(unsigned int); sized_type_for_promotion<3> ::type promoted_index_tester(long); sized_type_for_promotion<4> ::type promoted_index_tester(unsigned long); template<class T> struct promoted_index { static T testee; static const int value = sizeof(promoted_index_tester(+testee)); }; __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

4 7

Is there interest for a small portable garbage collection library?
by Achilleas Margaritis 27 May '06

27 May '06

Dear boost developers, Is there any interest for a small portable garbage collection library? I have written such a library to use for my personal projects and I would like to share it with the community, as I think it is very helpful in developing certain kinds of applications. Features: -small (under 2000 LOC) -portable (no lowlevel OS/CPU/compiler-specific features) -precise mark & sweep collection -no dependencies; headers only -weak pointers -garbage-collected arrays -customizable collector with user-defined finalizer and mark functions -wrapper class for external non-garbage-collected classes -thread-based collection "Thread-based collection" means the library can only support thread-specific collectors. I did not want to introduce dependencies on specific threading libraries, as they vary from project to project (and if many threads compete for the same collector, applications tend to slow down). The library, as it is, provides a collector for the main thread, but the function that returns the current collector can be modified to return a thread-specific collector (using thread-specific-storage). I have posted an article on www.codeproject.com about it along with the source code, code documentation and unit tests. It can be found here: http://www.codeproject.com/useritems/libgc.asp I have also posted a file release on sourceforge: http://sourceforge.net/projects/libgc I have tested the code (so far) in vc++ 7.1, but I think it will work on GCC and other compilers, as it does not use any O/S-specific and compiler-specific features. Even if boost devs find no value in it, I would appreciate any feedback. Thanks a lot for your attention. Achilleas Margaritis

4 7

SOC Project List Now Online
by Jeff Garland 27 May '06

27 May '06

The list of projects is now online at: http://code.google.com/soc/boost/about.html Not really much detail there, but hopefully more coming soon... Jeff

9 11

Re: [boost] Problem with range?
by Martin Bonner 27 May '06

27 May '06

Thomas Witt Sent: > Neal Becker wrote: >> >> With STL we can do: >> int x; >> std::copy (&x, &x+1, output); > > Really? Did you mean to say int x[1]; ? > Surely not. That would be: int x[1]; std::copy( x, x+1, output ); What Neal wrote is perfectly legal. See 5.7 para 4 in the standard: "For the purposes of these operators [+ and -], a pointer to a nonarray object behaves the same as a pointer to the first element of an array of length one with the type of the object as its element type." -- Martin Bonner Martin.Bonner(a)Pitechnology.com Pi Technology, Milton Hall, Ely Road, Milton, Cambridge, CB4 6WZ, ENGLAND Tel: +44 (0)1223 203894

2 1

Re: [boost] Infinite precision integer draft
by Martin Bonner 26 May '06

26 May '06

From: Maarten Kronenburg > To my knowledge (Aside: I think you mean by that "I am not aware of any evidence to contradict the suggestion that ...". If that is what you mean, "As far as I know ..." would be more idiomatic. Particularly as there is an English idiom "To my *certain* knowledge" which means "I know with absolute certainty that ...". N.B. Your English is already a whole heap better than my Dutch!) > the base type int does not > have a +0 and a -0. > So therefore I see no reason to specify this. C++ can be implemented on a sign and magnitude machine (I believe there are still some around). On such a machine an int can be either +0 and -0. (However I'm pretty sure the standard requires that those two values compare equal) > Derivation is a part of C++, so in my opinion > users must be able to derive from class integer > to make an integer with special properties. I think that would be a mistake. > How would you explain to a user that whatever > he/she does with integer, derivation is not an option, > because the destructor does not happen to be virtual. > This is what it boils down to in the end. They should use containment. > Once again I argue that an unsigned integer is an integer, > and a modular integer is an integer. > An unsigned int is actually a modular int with modulus 2^32. Actually the modulus is not specified. It cannot be smaller than 2^32, but it could be larger. -- Martin Bonner Martin.Bonner(a)Pitechnology.com Pi Technology, Milton Hall, Ely Road, Milton, Cambridge, CB4 6WZ, ENGLAND Tel: +44 (0)1223 203894

2 1

[pqs]Review dates confirmed
by Andy Little 26 May '06

26 May '06

Hi All, the proposed Boost.PQS (Physical Quantities System) library facilitates modelling physical quantities in C++ programs. The Formal Review dates for the review of PQS have now been confirmed... The PQS Formal Review will run from Wednesday 31st of May to Friday 9th of June. pqs_3_1_0 is the Official Boost Review version of PQS. (Please note that pqs_3_1_0 has a bug in one header. This can be either patched according to the latest release notes, "relnotes.txt" available from http://tinyurl.com/7m5l8 or alternatively simply download pqs_3_1_1, which is basically equivalent to pqs_3_1_0, except the bug has been fixed there. Suggestion: Composing a review of the PQS library could be a great way to relax! regards Andy Little

1 0