March 2010 - Boost - lists.preview.boost.org

Willing to Join GSOC 2010
by Xin Shuai 24 Mar '10

24 Mar '10

Hi: I'm interested in your Boost.Graph project. I think my current graph theory related research as well as C++ and python programming experiences makes me fit for the project. I'm wondering who's the mentor for the Boost.Graph project and I hope he/she can give me useful suggestions about the project. -- Xin(David) Shuai PhD of Complex System in School of Informatics & Computing Indiana University Bloomington 812-606-8019 The way to success is to do as much as important things, and as less as unimportant things, as you can...

2 1

[GSoC]I am interested in Boost.Graph
by doraemonok 24 Mar '10

24 Mar '10

My name is Rui Diao. I am a student of Graduate University of Chinese Academy of Sciences. My major is Computational Mathematics. I have never contributed to any open software project, but only spread the open source spirit to the people around me and join some open source groups. The only contribution of my source file is developing a Client of IP Gateway of My University by Qt open source version. Contributing to Boost C++ Libraries is an interesting and challenging job, and also, of course, fantastic. I am interested in one project of Boost.Graph : Graph Connectives. I like Graph Theory very much. I have experiences on coding (and teaching the friends in open software association of my university) some algorithms about Graph Theory: Dijkstra, Floyd, Bellman Ford, Ford Fulkerson, SAP, Kuhn-Munkres, etc. I have learned a lot in "How to finish an excellent project" during the internship of Microsoft Research Asia. Moreover, I am familiar with Mathematica which has similar function with Boost.Graph. I am very familiar with STL, but seldom use Boost. I just have experiences on using Boost.Regex before. I have checked out the source code by Subversion and began to read the part of Boost.Graph recently. The Boost.Graph library is missing connectives which contains Graph Join, Graph Union, Graph Product, Graph Intersection, Graph Difference, Graph Sum, Graph Product, Line Graph, etc. I think the algorithms may be not too difficult to me, but how to code with "Boost Style" (of course not only the code style) is a problem. I will try my best to finish the challenging work. I think may be there is some mentor can give me some advices or suggestions. Thank you all. Rui Diao

2 1

[log] Boost.Log formal review
by barend 24 Mar '10

24 Mar '10

Hi, This is my review of Boost.Log submitted by Andrey Semashev. > Please explicitly state in your review whether > the library should be accepted. I've been doubting about this for a few days. When I started my review, by trying to use the library, I discovered that the library is not header-only. I had thought that a logging library should be lightweight and had therefore expected only headerfiles. I then tried the library from John Torjo from 2007 but it is not header- only neither, so no reason to favour that one (I didn't dive into it further). Because of some messages on the list, a.o.: > I think that any reasonable log library must be compiled > to it's own lib due to the intermodule singleton requirement I realized that it is indeed useful to have a non-header-only option as well. The answers of Andrey were satisfactory, a.o.: > If we have a list of places in Boost where logging may be applied, > perhaps we could compile a set of requirements for such a lightweight > wrapper. When we have it, it would be easier to tell, whether Boost.Log > fits the role or not, or whether the wrapper should be part of it or a > separate submission. I'd be interested to hear from the library authors > on this topic. And I found in the Boost.Log documentation, TODO-list: > Think over a header-only configuration. Perhaps, with a reduced > functionality. So, good. Then I started the real review and I started to like the library, it is definitely useful, is configurable, trivial logging is quite simple. It also contains very useful non-trivial features such as writing to the Windows Event Log and alarm critical states as balloon tips. So I finally decided to give my YES vote, contingent on the condition that a lightweight header-only configuration, e.g. single-threaded, single-module, will be available for library writers. Some more about this below. --------------------------------------------------------------------- > What is your evaluation of the design? I like the options, the trivial logging, multiple back-ends, modularity and extensibility. I didn't have a detailed look to the design but in general it looks good to me. > What is your evaluation of the implementation? I didn't have a detailed look either. I glanced through the sources and the implementation is looking neat. I wonder why the class "basic_slim_string" is necessary, to me it sounds very strange why a task as logging needs an own string implementation. However, I didn't have the time to dive into it and didn't ask questions on this on the list. > What is your evaluation of the documentation? The documentation is good. Pages like design overview, tutorial, installation are easy to follow. The tutorial pages however contain some omissions and you have to refer to the examples to see what is missing. I mean here the chapter "Trivial logging with filters" which seems to give a complete example, but misses the namespace aliases, two of them would have been enough. > What is your evaluation of the potential usefulness of the library? Very useful, definitely, for applications and for libraries. > Did you try to use the library? With what compiler? Did you have any > problems? Yes, I used it using Visual C++ 2008 Express Edition. Because Boost.Log is not header-only, and I don't have the Boost libs compiled on my machine (don't ask me why, I never have them), I had several issues which other reviewers probably didn't run into. The defines _CRT_SECURE_NO_WARNINGS and _SCL_SECURE_NO_WARNINGS were necessary to surpress M$ warnings. The define BOOST_ALL_NO_LIB is of course necessary to build it like I did, and it works for Boost.Log. The compiler complained about a missing simple_event_log.h by event_log_backend.cpp, so I drafted it by hand, but I later saw that it is documented and you should neatly use the message compiler, so OK. After adding some Boost source files from thread, filesystem, regex and system it worked, datetime was not necessary. A complete build (including all mentioned sources) costs about 2 minutes. The tutorial did work (see message above on filtering, adding namespace aliasses was necessary) and I experimented with some things, many options seems very useful to me such as the log format, automatic date time, log file rotation, log file max size, etc. etc. I like the streaming feature (so for me printf is not necessary). I have some reservations about the compiler support: GCC 3.4.5, the default on Linux and MinGW, is not supported and for a task as logging this seems to me as a disadvantage. > How much effort did you put into your evaluation? A glance? A quick > reading? In-depth study? About five to six hours spread over some days. > Are you knowledgeable about the problem domain? Yes, I always need logging. In our library at Geodan we built logging as a singleton in 1995. We rewrote it to a logging-over-DLL's in 2000, but these two were not 10% as sophisticated and complete as this logging library is. So yes, these things were not header-only neither. --------------------------------------------------------------------- So back on the logging for libraries. If Boost libraries which are header-only will start to use this library, without protection, thousands and thousands of project files and makefiles worldwide will be broken. This answer: > If an existing header-only library starts using the Boost.Log library > we'll need to update the makefiles. This is a reasonable cost (not from Andrey)is therefore not satisfactory to me. This answer: > Which probably means that the decision is with the author(s) of each > individual library. (from the review manager) surprised me a bit, I think that in a library under review such an important thing may certainly be discussed, per library. However, as said, it is not a reason to vote no, because the rest of the library is quite good. A small macro (sorry) wrapper would already do most of the trick, and avoid breaking project files: #if defined(BOOST_ALL_LOG) || defined(BOOST_GEOMETRY_LOG) #define BOOST_LOG_GEOMETRY(o, s) BOOST_LOG_TRIVIAL(o) << s #else #define BOOST_LOG_GEOMETRY(o, s) #endif Where it can be called in libraries as: BOOST_LOG_GEOMETRY(info, "Log line with a value " << 3); This is inspired on the general define "BOOST_ALL_NO_LIB" and library- specific define "BOOST_LOG_NO_LIB". This small wrapper would really be useful, at least for us, and replace the #ifdef BOOST_GEOMETRY_DEBUG lines which are currently in our Boost.Geometry library. This is zero cost when one of the macro's BOOST_ALL_LOG or BOOST_GEOMETRY_LOG is not defined (by default it is not), and contains much functionality of Boost.Log as filtering, rotation, etc, when one of these macros is defined. Of course it can be enhanced or there are better solutions. Anyway, a lightweight solution will not be difficult. --------------------------------------------------------------------- So thanks Andrey for submitting this library, thanks Volodya for managing the review, and I hope it will be accepted with a lightweight header-only configuration included. Regards, Barend

2 1

interprocess
by Jan Stetka 24 Mar '10

24 Mar '10

Has anyone else noticed that a named_mutex gets destroyed after a windows xp computer goes into hibernate?

4 3

Problem regarding downloading Boost 1.42.0
by Wasik Mursalin Rushafi 24 Mar '10

24 Mar '10

Hi, I am trying to download the latest boost library from here<http://sourceforge.net/projects/boost/files/boost/1.42.0/>but failing. is there any problem? isn't it available or I am the only one facing this problem? -- Rushafi

2 1

[log] cygwin gcc-4.3.2 compilation/run-time errors
by vicente.botet 23 Mar '10

23 Mar '10

Hi Andrey, I have found thre kind of errors while compiling Bost.Log with cygwin gcc-4.3.2 * WHAR related * SYSLOG related Do you know if we can install syslog on cygwin? * SETTING related (reading from a configuration file) I have not yet analyzed the SETTING errors. I will see later. I have prefixed any modified line with the BUG WCHAR/SYSLOG/SETTING on the attached patches. After all this I have reached to compile the whole library with tests and examples. Only 2 test passed and There are yet some link warnings that could be suspect and a lot of runtime errors (see below) It would be great if you can apply some of the modification on the attached patch (of course not the Jamfiles) Best, _____________________ Vicente Juan Botet Escribá gcc.link ../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.exe Info: resolving vtable for std::exception by linking to __imp___ZTVSt9exception (auto-import) Info: resolving vtable for std::bad_cast by linking to __imp___ZTVSt8bad_cast (auto-import) Info: resolving std::cout by linking to __imp___ZSt4cout (auto-import) Info: resolving std::cerr by linking to __imp___ZSt4cerr (auto-import) Info: resolving typeinfo for std::runtime_error by linking to __imp___ZTISt13runtime_error (auto-import) Info: resolving std::ios_base::in by linking to __imp___ZNSt8ios_base2inE (auto-import) Info: resolving std::ios_base::out by linking to __imp___ZNSt8ios_base3outE (auto-import) Info: resolving typeinfo for std::bad_alloc by linking to __imp___ZTISt9bad_alloc (auto-import) Info: resolving typeinfo for std::bad_cast by linking to __imp___ZTISt8bad_cast (auto-import) Info: resolving typeinfo for std::bad_typeid by linking to __imp___ZTISt10bad_typeid (auto-import) Info: resolving typeinfo for std::bad_exception by linking to __imp___ZTISt13bad_exception (auto-import) Info: resolving typeinfo for std::logic_error by linking to __imp___ZTISt11logic_error (auto-import) Info: resolving typeinfo for std::exception by linking to __imp___ZTISt9exception (auto-import) Info: resolving typeinfo for std::basic_ostringstream<char, std::char_traits<char>, std::allocator<char> > by linking to __imp___ZTISt19basic_ostringstreamIcSt11char_traitsIcESaIcEE (auto-import) Info: resolving typeinfo for std::basic_ostream<char, std::char_traits<char> > by linking to __imp___ZTISo (auto-import/usr/bin/ld: warning: auto-importing has been activated without --enable-auto-import specified on the command line. This should work unless it involves constant data structures referencing symbols from auto-imported DLLs.) testing.capture-output ../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.run ====== BEGIN OUTPUT ====== EXIT STATUS: 53 ====== END OUTPUT ====== PATH="//bin://lib://lib32://lib64:/boost_1_41_0/bin.v2/libs/date_time/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/filesystem/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/log/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/regex/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/system/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/test/build/gcc-4.3.2/debug/threading-multi:/boost_1_41_0/bin.v2/libs/thread/build/gcc-4.3.2/debug/threading-multi:$PATH" export PATH "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.exe" > "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.output" 2>&1 status=$? echo >> "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.output" echo EXIT STATUS: $status >> "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.output" if test $status -eq 0 ; then cp "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.output" "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.run" fi verbose=0 if test $status -ne 0 ; then verbose=1 fi if test $verbose -eq 1 ; then echo ====== BEGIN OUTPUT ====== cat "../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.output" echo ====== END OUTPUT ====== fi exit $status ...failed testing.capture-output ../../../bin.v2/libs/log/test/attr_attribute_values_view.test/gcc-4.3.2/debug/threading-multi/attr_attribute_values_view.run...

1 0

Re: [boost] 5 Observations - My experience with the boost libraries
by Brian Wood 23 Mar '10

23 Mar '10

Tom Brinkman writes: >> Exceptions are a fact of C++ life. You can't avoid them if for no other >> reason than std::bad_alloc, unless you don't allocate from the free store. > > Not true, there are no-throw versions of std::new. > > In any event, my point is that c++ exception handeling should be optional. > Boost libraries need to be updated to reflect this. I used to support both return codes and exceptions. Eventually I gave up supporting return codes, because attempting to support both was quite difficult. Dropping support for return codes was kind of a difficult decision for me -- but after not being able to think of another project that supports both, I decided to only support exceptions. Since making that decision, I feel I'm able to make faster progress in other areas. Do you know of any libraries that support both return codes and exceptions? -- Brian Wood Ebenezer Enterprises http://webEbenezer.net (651) 251-9384

1 0

Re: [boost] GSOC 2010 Boost.Graph
by Chinmaya Poswalia 23 Mar '10

23 Mar '10

Please comment on the following topics Hierarchical Layout Algorithms : 1) what is being implemented in spring based layout or what are the files(headers files) in the boost. Computer Vision Algorithm : 1) suggest me some examples regarding the "Implementing a specialized graph type for grids would also be needed in order to avoid storing the graph topology explicitly". I am not getting the point correctly.

2 1

[Review] ITL review extended until March 7th
by Hartmut Kaiser 23 Mar '10

23 Mar '10

Hey all, After getting the approval from our review wizards we decided to extend the review period of Joachim Faulhabers ITL library until March 7th, 2010. That should give everybody some more time to finish their reviews. Regards Hartmut --------------- Meet me at BoostCon www.boostcon.com

4 4

[Boost.Locale] Questions (was New Release)]
by Alexander Lamaison 23 Mar '10

23 Mar '10

Artyom wrote: >> I sent this to the list via Gmane but >> its stuck in the moderation queue so >> I'm sending it to you personally: >> > You probably should register to boost mailing list directly? > I did that too but it said because it was my first post to the list (I usually hang out in boost.user not .devel) it had to be moderated. It's been days now :( >> I compile my code with MSVC8: >> > > I hadn't tested this library with MSVC8 (2005), I would be glad if > you can build it and run unit-test (if you using nmake then > run "nmake test") If you are working with solution there should be > an option. > I did this and everything passed. Though when I run examples on the command line the output includes much that isn't readable. Is this because the terminal doesn't handle the codepages correctly? > So I would know that MSVC 2005 is fine as well. Once you run the > test, can you please send me: your os version and ICU Version so I can > update the tested platforms list? > I'm using Vista 32-bit with ICU 4.2.1 I've just started internationalising my project (Swish http://swish.sourceforge.net) using Boost.Locale so I may have many questions for you as I go along. I hope this will helps Boost.Locale as much as it helps me. Firstly, I'm trying to decide which format() to use: boost::format or boost::locale::format. If I don't use currencies or dates and always use the global locale, are there any advantages to boost::locale::format? The tutorial says of boost::format "It knows nothing about new Boost.Locale manipulators". I don't understand what this means. Which new manipulators? I notice that boost::locale::format has a constructor that takes a message as an argument and then does something internally that depends on the value of _translate but I don't understand why. boost::format can be used like this: boost::format(translate("Id string %1%")) % arg so what does the extra boost::locale::format constructor allow that this doesn't? Also I make gettext enforce correct translation of format strings by passing --boost to xgettext. This reports an error if the format tokens, %1% etc., don't match in the original and translated strings. Changing these to {1} etc. in boost::locale::format means this no longer works. What was the tecnical motivation for this change? My second question relates to the directory structure for the .mo files and applies to gettext as much as to boost::locale (I'm new to gettext too). Why do they expect to find the files at lang_ctry/LC_MESSAGES/proj.mo? This makes sense when translations for many projects are stored in a common location such as |/usr/share/locale/ but on Windows this wouldn't be likely. I plan to store my translation in my program's installation directory and wish I could just put them in a flat directory: translations/ar.mo translations/en_GB.mo etc. Is it possible to make the lookup fallback to this scheme or even make the scheme customisable? Many thanks for this excellent library. Alex |

2 1