- Boost-users - lists.boost.org

[program_options] unknown options in config files
by Sebastian Weber 20 Feb '07

20 Feb '07

Hello! I'm using program_options to parse command line options and configuration files. I absolutley need the unknown option feature. For the command line options this works fine. However, the configuration file parsing always throws an option unknown exception when I parse cfg-files with non-registered options. Right right I circumvent this by catching the corresponding exception. But this is not very nice and I can as well not be sure that all options have been parsed in the cfg-file, depending upon the first occurence of an unknown option, right? So, is there a possibility to process unknown options in cfg-files? Thanks a lot in advance. Greetings, Sebastian

1 0

Boost in embedded systems
by Peter Poulsen 20 Feb '07

20 Feb '07

I know this sounds insane, but our developers are very happy with boost (who can blame them :-)), and want to have boost available on our embedded systems. We only have 64 mb of flash memory available for the entire system (OS, hardware monitoring and applications) and I simply cannot get boost down to a reasonable size. Am I trying to do an impossible task, or can somebody give me some good advice on how to build boost so it has a more reasonable size? Of course what should be done is removing some of the features, and I can probably negotiate with developers about what features we can provide and which have to go. Which one should I try to get rid of first? /peter -- A: Because it messes up the order in which people normally read text. Q: Why is it such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?

5 4

Updated date_time_zonespec.csv for 2007?
by Brendon Oliver 20 Feb '07

20 Feb '07

Does anybody have a date_time_zonespec.csv file updated for the US DST changes this year? I've googled the mailing lists & looked at sourceforge (I see there was a bug raised 5 Feb, but no responses since). I also downloaded a 1.34 alpha tarball but there's no date_time_zonespec.csv file in there :-( . I've looked at updating the file myself (the syntax is pretty straight-forward), I'm just worried about missing particular zones. Cheers, - Brendon. -- Conversation, n.: A vocal competition in which the one who is catching his breath is called the listener. 10:19:19 up 61 days, 19:45, 2 users, load average: 0.34, 0.26, 0.25

2 3

[signals] thread_safe_signals feature-complete
by Frank Mori Hess 20 Feb '07

20 Feb '07

Hi all, I've finished thread_safe_signals. To get thread safety you have to include an additional header file thread_safe_signals/multi_threaded.hpp and set the last template parameter to boost::signals::multi_threaded. I've done a little benchmarking. With thread-safety off and connecting 10 slots, it appears to have about the same overhead as cvs boost.signals (which is much less than the latest release of boost.signals). When 1 slot is connected, it has more overhead, a little less than a factor of 2 when running a single empty slot. Turning thread-safety on increases overhead by about 50% for the 1 slot case and 100% for the 10 slot case. It is able to invoke slots concurrently though, which should be a win if the slot is non-trivial. -- Frank

4 38

any boost library for referring a portion of the container ?
by abir basak 19 Feb '07

19 Feb '07

Hi, I am having a few containers whose some portion I want to refer to another class (sometime I need to refer to a continuous portion, while sometime need a non-continuous portion) . At present I have written all of these myself. But wondering if a few portion can be substituted by some boost library. Just to make the concept clear, I am elaborating the structure and why it is done in that way. In the program it takes the user handwriting (as entered using digital pen etc) and analysis them. as user writes, I register points (x,y coord) and put them in an deque (in actual case I am using boost::circular_buffer_space_optimized ). Now, a second deque (or other container) stores another structure Trace, which has 2 index pointing start & end of that continuous writing until pen up occurs. These have to be 2 index rather than iterator, as iterator may get (and usually gets ) invalidated while more points are added in the point deque. Thus Trace refers deque<Point> Now from the trace structure I return the points for the trace (boost::sub_range ) using the 2 index and the deque (i.e from deque<Point> get iterator, add index as offset , make sub_range and return). Now these are the continues portions , referred by 2 index. Next, some such things can be even not continuous. In that case I store all index to which it points (in a vector ) and return a sub_range kind of structure (named index_range ) which takes these index and a container, and iterates over it jumping by those specific offsets. Example can be a Character class which refers some of those traces , not necessarily continuous (like user writes a dot much later than the i , thus Character i contains index of the vertical stroke of the i & dot ) Similarly more hierarchy are constructed, like Trace => Page, and Character => Word => Line => Paragraph => Page etc. Most of them contains 2 index and with the help of the container returns a boost sub_range , while some contains a vector of index (size_t) and returns index_range (created by me) . So far I found it is the ONLY way to create a level of hierarchy/ view from different set of dynamic data. (The other can be storing everything as a pointer in the container, and creating by new, in that case no problem arises as the objects are not movable, but in C++ it is quite bad solution) A Graphical representation is like, =============== deque<Point> ===================== <=remove old data => add new data | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | =============== deque<Trace> ==================== | 1 | 2 | =============== deque<Page> ===================== These are for continuous data (i.e 2 index struct & boost::sub_range ) ================ deque<Trace> ==================== |1,3,4| 2 | 5,6 | 7,9 | 8 | ... ================== deque<Character> =============== | 1,2,4, 5 | 3 | .... ================== deque<Word> ===================== | 1 | 2 | 3 | .... (assuming past a line user doesn't correct word, they can be made a pair of index ! ) ================== deque<Line> ===================== .... ================== deque<Paragraph> ================ etc... My question , 1) had anyone faced these kind of design , where a class referrers a portion of STL container ? and multiple view from same set of data exists (otherwise I could had Trace contains the points , instead of having index of the portion, if Trace is the only view for the points ) 2) Any boost library / other library is there to help these kind of design ? I had posted a same kind of topic in C++ newsgroup also, but it seems their people are more concerned with answering C++ standard based question . I think, boost may be the better news group to ask this question, as boosters work in a broader area. Thanks to all, esp to those who share their thought on it ... abir -- Abir Basak, Member IEEE B. Tech, IIT Kharagpur email: abir(a)abirbasak.com homepage: www.abirbasak.com

2 5

[boost::lexical_cast] problems casting hex numbers
by Ovanes Markarian 19 Feb '07

19 Feb '07

Thanks a lot for lexical_cast it is a very useful function. Currenlty I face a problem with it and I think it would be useful addon ;) Unfortunately it can not be used (or I was unable to figure out how) to cast hex or octal numbers from strings to numeric. If it is not possible, are there some concerns which disallow this? the expression: lexical_cast<size_t>("0x015e") throws a bad_lexical_cast With Kind Regards, Ovanes

1 0

BoostCon 2007 update
by David Abrahams 19 Feb '07

19 Feb '07

A couple of quick items: 1. The conference format has changed. For many practical reasons, the "two 3-day tracks with one day of overlap on Wednesday" format proved unworkable. Among those: * Many (most) sessions don't fall neatly into the library-user/library-developer track distinction. We see this as a good thing, as it means we have more material that's of broad interest. * It made the hotel reservation pattern very difficult; we were afraid many people would be shut out of rooms at the conference hotel and also that the hotel would not be able sell the entire reserved room block, which could cost the (non-profit) conference a lot of money, as we had to agree to be liable for unsold room nights. Instead, we're running two tracks concurrently all week. We thought attendees should be prepared for that change. 2. We expect the new boostcon website will go live and registration for the conference will open in the next couple of days. We're just clearing away a few final details. In the meantime, the existing http://www.boostcon.com has the old information. Regards, and see you at BoostCon, -- Dave Abrahams Boost Consulting www.boost-consulting.com

1 0

Boost::Date_Time -New IO Tutorial
by Steve Johns 19 Feb '07

19 Feb '07

Jeff, Thanks for your answers to my earlier posts. This is a great library, and I'm enjoying learning it. :^) I notice here: http://www.crystalclearsoftware.com/libraries/date_time/release_1_33/date_t… that a tutorial on the new IO is supposed to be found here: libs/date_time/examples/tutorial This dir is not part of my Boost distro. Is there a URL I can use to access this?

2 1

Problem in building boost 1_33_1 on solaris 5.9
by Shobha_Hubballi＠i2.com 19 Feb '07

19 Feb '07

Hi, I am trying to build boost1_33_1 on solaris 2.9 with CC 5.5 compiler. Libraries like filesystem, serialization, program_options are not built. But other libraries like thread are built perfectly. The libraries which I am not able build I am getting following same error for all libraries /data/tdm/boost_1_33_1/boost/iterator/iterator_facade.hpp", line 838: Error: Could not find a match for boost::iterator_core_access::equal<boost::iterator_core_access::Facade1, boost::iterator_core_access::Facade2>(const boost::filesystem::path::iterator, const boost::filesystem::path::iterator, boost::is_convertible<boost::filesystem::path::iterator, boost::filesystem::path::iterator>). "/opt3/SUNWspro/prod/include/CC/Cstd/./algorithm.cc", line 2370: Where: While instantiating "boost::operator!=<boost::filesystem::path::iterator, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>, boost::bidirectional_traversal_tag, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, int, boost::filesystem::path::iterator, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>, boost::bidirectional_traversal_tag, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, int>(const boost::iterator_facade<boost::filesystem::path::iterator, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>, boost::bidirectional_traversal_tag, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, int>&, const boost::iterator_facade<boost::filesystem::path::iterator, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>, boost::bidirectional_traversal_tag, const std::basic_string<char, std::char_traits<char>, std::allocator<char>>&, int>&)". Please tell me how to solve this problem Thanks

2 3

[Program option] Parsing string to vector<string>
by Seweryn Habdank-Wojewódzki 19 Feb '07

19 Feb '07

Dear all I have question. to the type of object stored in variables_map. In the example multiple_sources.cpp There is line: config.add_options() //... ("include-path,I", po::value< vector<string> >()->composing(), "include path") ; But independently of the case all time result vector has size = 1, so in fact above is equivalent: config.add_options() //... ("include-path,I", po::value< string >()->composing(), "include path") ; And still there is some parsing job to do. How can it be improved? I have tested this example using config file like below: # # Comment out this line to use hard-coded default value of 10 # optimization = 1 include-path = /opt /usr /home/ Regards. -- |\/\/| Seweryn Habdank-Wojewódzki `\/\/'

1 1