- Boost - lists.boost.org

Re: [boost] [build][msvc] create PDBs
by Ahmed Charles 24 Jul '13

24 Jul '13

The size difference is LTCG, not debug info. ________________________________ From: Daniela Engert Sent: 7/23/2013 9:40 AM To: boost(a)lists.boost.org Subject: Re: [boost] [build][msvc] create PDBs Am 23.07.2013 17:47, schrieb Andrey Semashev: > On Tue, Jul 23, 2013 at 7:39 PM, Daniela Engert <dani(a)ngrt.de> wrote: >> >> libboost_log_setup-vc120-mt-1_54.lib 188.671.120 >> libboost_log_setup-vc120-mt-1_54.pdb 132.616.192 >> libboost_log_setup-vc120-mt-gd-1_54.lib 220.099.308 >> libboost_log_setup-vc120-mt-gd-1_54.pdb 132.837.376 >> libboost_log-vc120-mt-1_54.lib 53.565.446 >> libboost_log-vc120-mt-1_54.pdb 7.942.144 >> libboost_log-vc120-mt-gd-1_54.lib 32.516.250 >> libboost_log-vc120-mt-gd-1_54.pdb 8.114.176 >> > > It looks like static libs contain debug info anyway as they are always > larger than the generated pdb. In many cases, your observed size relationship is true, but not in all of them. E.g.: libboost_exception-vc120-mt-1_54.lib 53.642 libboost_exception-vc120-mt-1_54.pdb 102.400 libboost_exception-vc120-mt-gd-1_54.lib 11.440 libboost_exception-vc120-mt-gd-1_54.pdb 102.400 To exhibit the size advantage of PDBs over 'old-style' debug info embedded into static libs, I've rebuilt Boost.Log with link=static debug-store=object. This is the result: libboost_log_setup-vc120-mt-1_54.lib 343.234.460 libboost_log_setup-vc120-mt-gd-1_54.lib 374.488.888 libboost_log-vc120-mt-1_54.lib 79.564.296 libboost_log-vc120-mt-gd-1_54.lib 59.539.544 Ciao Dani _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

2 1

clang AST based static analysers - Booster's experience
by Niall Douglas 23 Jul '13

23 Jul '13

Dear Boost Devs, I was wondering what the experience of the list is with clang AST based static analysers? They need not be clang AST based of course, so long as they are based on the frontend of a real compiler rather than trying their own implementation based on regex or something even worse. For this list's information, here are the three clang AST based static analysers I know of: 1. Clang's static analyser http://clang-analyzer.llvm.org/ 2. OCLint http://oclint.org/ 3. The commercial tool CppDepend can now use a clang AST (http://blog.llvm.org/2013/04/static-analysis-tools-using-clang-in.html) Some quick notes of mine which may be of dubious accuracy: Clang's static analyser is the more mature tool, but according to http://code.google.com/p/chromium/wiki/ClangStaticAnalyzer which was updated Mar 2013 it feels its default bundled C++ tests is lacking. The real strength in the clang static analyser is for writing your own tests (http://clang-analyzer.llvm.org/checker_dev_manual.html) which enforce specific coding patterns or constraints (usually on others, and usually pre-commit with commit rejection if fail). OCLint is a newer tool, but appears to have fuller support for C++ (http://docs.oclint.org/en/dev/rules/index.html) though many of those rules appear to check C syntax only (which also covers Objective C). Similar to clang's static analyser, new custom rules are easy to write (http://docs.oclint.org/en/dev/internals/rules.html) This project, at least from the outside, seems to have more legs than clang's static analyser. CppDepend I have no experience of, but there is a reasonably positive review at http://www.asawicki.info/news_1511_cppdepend.html. One of the CppDepend developers talks about how they added clang's AST at http://www.codeproject.com/Articles/475254/ClangplusRocks. Niall --- Opinions expressed here are my own and do not necessarily represent those of BlackBerry Inc.

1 0

[build][msvc] create PDBs
by Daniela Engert 23 Jul '13

23 Jul '13

Hi list! For various reasons I always preferred to have debug info separate from the binaries. Unfortunately, emitting debug info to PDBs doesn't work so well, in particular when it comes to static libs and precompiled headers. Because of that I tried to rectify this and created a patch against msvc.jam. (see ticket #8889 https://svn.boost.org/trac/boost/ticket/8889) This (and a tiny script to collect all generated PDBs into the library directory) got me there: boost_log_setup-vc120-mt-1_54.dll 1.158.656 boost_log_setup-vc120-mt-1_54.lib 24.244 boost_log_setup-vc120-mt-1_54.pdb 123.055.104 boost_log_setup-vc120-mt-gd-1_54.dll 3.206.144 boost_log_setup-vc120-mt-gd-1_54.lib 24.408 boost_log_setup-vc120-mt-gd-1_54.pdb 219.220.992 boost_log-vc120-mt-1_54.dll 486.400 boost_log-vc120-mt-1_54.lib 241.174 boost_log-vc120-mt-1_54.pdb 10.857.472 boost_log-vc120-mt-gd-1_54.dll 1.071.616 boost_log-vc120-mt-gd-1_54.lib 242.278 boost_log-vc120-mt-gd-1_54.pdb 18.918.400 libboost_log_setup-vc120-mt-1_54.lib 188.671.120 libboost_log_setup-vc120-mt-1_54.pdb 132.616.192 libboost_log_setup-vc120-mt-gd-1_54.lib 220.099.308 libboost_log_setup-vc120-mt-gd-1_54.pdb 132.837.376 libboost_log-vc120-mt-1_54.lib 53.565.446 libboost_log-vc120-mt-1_54.pdb 7.942.144 libboost_log-vc120-mt-gd-1_54.lib 32.516.250 libboost_log-vc120-mt-gd-1_54.pdb 8.114.176 Boost.Log was chosen for exhibition only. All files were built with debug-symbols=on debug-store=database. Link-time code generation was enabled for all release builds, too. May be, someone finds this helpful. Ciao Dani

2 2

[format] doc fix
by Tim Blechmann 23 Jul '13

23 Jul '13

not sure, how well-maintained boost.format is these days ... any objection if i commit this documentation fix? cheers, tim

1 0

Fwd: Re: Announcing boost dependency analyzer
by Stephen Kelly 23 Jul '13

23 Jul '13

Jens Weller wrote: >> Unfortunately the Boost dependency graph is as huge a hairball as I was >> afraid :-/ > > Hi, thanks for the feedback. Yes, dependency graph for boost is quite > large... I think with boost in git and modules available, this is going to > improve. That won't happen automatically. The git conversion is separate to reorganizing to resolve the solvable interdepency issues: http://thread.gmane.org/gmane.comp.programming.tools.ryppl.devel/201/focus=… The salient point is: > By modularizing and making the interdependencies explicit, we hope to > encourage developers to refactor. But we can only solve one problem at a > time. So, if you want to work on resolving interdependencies, don't wait for the git conversion to be done, because it won't help, and do pick up your editor and dig, because only that will help. Thanks, Steve.

1 0

Regression tests report page cycling slow
by Joaquin M Lopez Munoz 23 Jul '13

23 Jul '13

Hi, the page http://www.boost.org/development/tests/trunk/developer/summary.html seems to update very unfrequently, every two days or longer. Can something be done to speed this up? Thank you in advance, Joaquín M López Muñoz Telefónica Digital

2 1

Requesting number crunchers for trial complex classes
by Daryle Walker 22 Jul '13

22 Jul '13

I just pushed out commit "aaeabbd3fcc04956c4fd3d3825f6202355ab9297" of my trial hypercomplex number classes at https://github.com/CTMacUser/Complex. The classes now have all the arithmetic operators. They already have the condition functions (norm and conj), norm functions (taxi (ex L1), abs, sup, and sgn), and equality operators. There is an output operator, but not input. There are constructors for: default (both garbage bits and all zeros), copy and move (automatically defined), component (single or list), cross version, smaller hypercomplex (single and list), and (head of) larger hypercomplex. Look at the unit tests for usage examples. I haven't done any utility or transcendental functions yet, so let me know if you need them before proceeding. I just want people who do hard-core number crunching with complex, quaternion, and/or octonion numbers to see if my classes are worth it, performance-wise. You can also use the array-of-hypercomplex to array-of-many-reals reinterpret_cast trick. Everything should have Doxygen comments. I run this code on GCC 4.7.1 off MinGW for 32-bit Windows 8 Pro, so tests on other compilers will be appreciated. (It probably can't run on Visual C++, but GCC 4.8 and Clang 3.x will probably work.) Note on octonions: a month ago I found out, while looking for online octonion calculators, that octonion (and above) multiplication isn't standardized. I'm using the same philosophy from Wikipedia, which is also used by the existing Boost.Octonion. (Actually, I found a bug in B.O.'s octonion times quaternion code, since that author manually lists all the sub-terms, and my code is more automatic. I had to do the multiplication by hand to check who had it right.) So if your octonion products are different, it may not be a bug. I've tried to make as many constructors/operators/functions constexpr as possible. (That's why no std::initializer_list.) That may help move data into ROM, but you have to force non-inline-ness when you need to test runtime performance. Daryle W.

1 0

[Locale] does generator("") deliver the system's default locale?
by Michael Schinner 22 Jul '13

22 Jul '13

boost 1.54.0 When I use boost::locale::generator and generator("") I get different results for the decimal point than using std::locale(""). Boths methods are supposted to deliver the system's default locale. Example: // windows 7, de_AT std::use_facet<std::numpunct<char> >(std::locale("")).decimal_point(); // delivers a , boost::locale::generator generator; std::use_facet<std::numpunct<char> >(generator("")).decimal_point(); // delivers a . What does generator("")) deliver? Is it wanted that a '.' is deliverd as decimal point for a german locale? Best regards, Michael

2 1

[system] Do we need this library with latest (almost C++0x) compilers?
by Tonko Juricic 22 Jul '13

22 Jul '13

Can somebody please explain what is missing in current compiler implementations? What is preventing other libraries to have a macro like: #ifdef _CPP11_THREAD_ #include <thread> #include <mutex> #include <condition_variable> #else #include <boost/thread.hpp> #include <boost/thread/mutex.hpp> #include <boost/thread/condition_variable.hpp> #endif also when it comes to error codes and conditions? Thanks, Tony

3 2

windows_shared_memory - does memory block exists?
by piotr.ns 21 Jul '13

21 Jul '13

Hi, Why native windows shared memory does not include any means to check for existance of memory block without throwing exceptions? A small function would be enough: // boost::interprocess::windows_shared_memory static bool exists(const std::string &a_path) { using namespace boost::interprocess; try { unsigned long file_map_access = winapi::page_readonly; unsigned long map_access = winapi::file_map_read; void *test_handle = winapi::open_file_mapping(map_access, a_path.c_str()); bool res = (test_handle != NULL); if(res) winapi::close_handle(test_handle); return res; } catch(...) { return false; } } Without this I have to use Boost internals (winapi::*) in my application. Would somebody consider including this code or something similar? Regards, Piotr

1 0