November 2004 - Boost - lists.boost.org

RE: [boost] Re: GUI Library Proposal for a Proposal
by Reece Dunn 08 Nov '04

08 Nov '04

Russell Hind wrote: >Edward Diener wrote: >> >>No !!!!! WxWidgets has almost nothing to do with modern C++. Let's no go >>there, or have anything to do with supporting any other GUI >>implementation. > >Also, wxWindows is a huge library. If we were to introduce something in >boost, I'd very much like it to be as small and light weight (as win32gui >is supposed to be), not a collection of multi-megabyte dlls required, just >to wrap win32 or what ever the underlying library is. I am in the process of writing a platform-independant GUI (available at: http://uk.geocities.com/msclrhd/gui/gui.zip) that aims to be small and lightweight. This code is stll in the early stages and currently targets Win32/64 with minimal Cocoa/PalmOS support. >If people want more powerful components, they can be added afterwards, but >the underlying library around core OS controls should be as small as >possible. In my library, I am working on: * gui::event_handler supporting lightweight "windowless" controls * gui::window supporting windows like the main frame, MDI child frames, etc. * gui::form (i.e. a frame with controls/layout set by an external designer referenced by ID) that includes dialog boxes in Win32. I do not have an implementation of gui::form yet. * gui::control supporting controls/widgets like buttons within a window or form These 4 classes should then make it possible to derive your own classes to support edit controls, property sheets, MDI windows, etc. I do not explicitly implement any of these extra UI components because they are platform dependant and will add to the complexity of the framework. This is especially true for things like "splitter" windows and floating/docking toolbars. Also, other components can be added at a later stage if necessary. As for the event processing mechanism, I have decided to use a std::map< event_id, void ( gui::event_handler::* )( gui::event * ) > that allows: class main_frame: public gui::window { public: void ev_destroy( gui::event * ); void EvDraw( gui::event * ); main_frame( gui::string name ): gui::window( ... ) { set_handler( WM_DESTROY, &main_frame::ev_destroy ); set_handler( WM_PAINT, &main_frame::EvDraw ); } }; At the moment, this does not support chaining of handlers (something on my todo list). The main event loop is controlled via gui::event_loop that does support filtering of events (e.g. processing keyboard shortcuts). There are also platform-independant versions of position (NSPoint, POINT, etc.), size (NSSize, PointType, etc.) and area (NSRect, RECT, etc.) that make use of properties allowing window positions to be handled in a platform-independant manner, e.g.: void main_frame::resizing( gui::event * ) { area a = get_client_area(); a.deflate( gui::size( 5, 5 )); pane -> move( a ); } At the moment, you need to register Win32/64 window classes using: gui::win::window_type frame_class( TEXT( "CxxGUIFrame" ), COLOR_BTNFACE ); but I am looking to abstract this. I am all for getting a group of people together from the boost mailing list interested in GUI and developing a general GUI framework. It may be useful to have at least one person from the C++ standards committee to get input from the standards people. Regards, Reece _________________________________________________________________ Want to block unwanted pop-ups? Download the free MSN Toolbar now! http://toolbar.msn.co.uk/

10 26

how to compile regression tests
by Roland Schwarz 08 Nov '04

08 Nov '04

Sorry for this possibly naive question. But I can't figure out how to compile the regression tests. So far I found cd'ing into a specific libs/test directory and invoking bjam on it compiles the test for a specific library. But how can I compile all? I cannot find a rule like stage or install in the root Jamfile. Isn't it possible to recurse through all directories? Then I cannot manage to get a results page using the description in http://www.boost-consulting.com/boost/more/regression.html Instead I get: > C:\Programme\boost>SET PYTHON_ROOT=C:\Programme\Python23 > > C:\Programme\boost>SET BOOST_ROOT=C:\Programme\boost > > C:\Programme\boost>SET TOOLS=vc-7_1 > > C:\Programme\boost>ECHO ON > > C:\Programme\boost>cd C:\Programme\boost\status > > C:\Programme\boost\status>bjam --dump-tests test 1>bjam.log 2>&1 > > C:\Programme\boost\status>REM start notepad bjam.log > > C:\Programme\boost\status>process_jam_log 0<bjam.log > Usage: bjam [bjam-args] | process_jam_log [--echo] [locate-root] > locate-root is the same as the bjam ALL_LOCATE_TARGET > parameter, if any. Default is boost-root. > boost_root: C:/Programme/boost/status/.. > locate_root: C:/Programme/boost/status/.. > no errors detected > > C:\Programme\boost\status>compiler_status C:\Programme\boost cs-win32.html > > **** exception(205): std::exception: > boost::filesystem::directory_iterator constructor: > "C:\Programme\boost\status\bin": > The system cannot find the path specified. > ******** errors detected; see standard output for details ******** > > C:\Programme\boost\status>rem Specify links file, even though it will > be overwritten, so report html includes links > > C:\Programme\boost\status>rem This works because the generated > bookmark names are the same regardless of other settings > > > C:\Programme\boost\status>compiler_status --ignore-pass --no-warn > C:\Programme\boost cs-win32-fail.html cs-win32-links.h > tml > > **** exception(205): std::exception: > boost::filesystem::directory_iterator constructor: > "C:\Programme\boost\status\bin": > The system cannot find the path specified. > ******** errors detected; see standard output for details ******** > > C:\Programme\boost\status>compiler_status --ignore-pass > C:\Programme\boost cs-win32-warn-or-fail.html cs-win32-links.htm > l > > **** exception(205): std::exception: > boost::filesystem::directory_iterator constructor: > "C:\Programme\boost\status\bin": > The system cannot find the path specified. > ******** errors detected; see standard output for details ******** > > C:\Programme\boost\status>compiler_status C:\Programme\boost > cs-win32-full.html cs-win32-links.html > > **** exception(205): std::exception: > boost::filesystem::directory_iterator constructor: > "C:\Programme\boost\status\bin": > The system cannot find the path specified. > ******** errors detected; see standard output for details ******** > > C:\Programme\boost\status> I am using vc7.1 on W2 with the latest CVS MAIN. I currently assume this should work under windows, doesn't it? Thank you for help, Roland

1 0

build / install in 1.32
by John Maddock 07 Nov '04

07 Nov '04

Should the following messages be generated during an install? They are rather alarming IMO, there are a huge number of them, and I predict will generate queries along the lines of "is this right, did my build succeed?" HardLink stage\lib\boost_thread-vc71-mt.lib Could Not Find c:\data\boost\release\boost\stage\lib\boost_thread-vc71-mt.lib HardLink stage\lib\libboost_thread-vc71-mt-sgd.lib Could Not Find c:\data\boost\release\boost\stage\lib\libboost_thread-vc71-mt-sgd.lib HardLink stage\lib\libboost_thread-vc71-mt-s.lib Could Not Find c:\data\boost\release\boost\stage\lib\libboost_thread-vc71-mt-s.lib Thanks, John.

3 2

Getting the reference out of boost::any
by Mikhail Glushenkov 07 Nov '04

07 Nov '04

I became interested in subj and quickly found a discussion on boost mailing list dated Jan 2003 (http://lists.boost.org/MailArchives/boost/msg42284.php) where the majority seemed to agree on changing current semantics of boost::any_cast to allow getting the reference to held object. However, these changes never made it to the CVS. Was it by intent or was this issue banally forgotten?

1 0

Unicode strings update
by Erik Wien 06 Nov '04

06 Nov '04

Hi. I just thought I would post a little update on what's happpening with the Unicode library I started a thread about a little while ago. (I apologize for not replying as much as I should have done in that thread, but I have been completely buried in work lately, and simply haven't had the time. =\ ) My college (HiG) has now basically approved the project as a bachelors thesis (at least for the time being), and it is looking very likely that we will go through with it. There's one little snag though. HiG would like to have one person from the boost community functioning as a "contact person" for the project. I'm not sure exactly what kind of workload that imply for such a person (I will try to find out very soon), but I am fairly certain it will just be a matter of answering a few e-mails from them from time to time. After all, the main development will be done in collaboration with the community as a whole, and not that single person. What I basically want to know is, would anyone out there like to take on that resposibillity? I don't think this is an absolute requirement for final approval, but it sure would help! =) Let me just finish off by thanking for all comments you guys made in my first thread. I'm in the process of compiling a list of all the things you pointed out there, and I will most certainly take everything said into concideration when the development (hopefully) gets on the way. As always... Any comments are welcome! ;) Regards - Erik

7 12

usage of uniform_int
by Masaki Nomura 06 Nov '04

06 Nov '04

Dear all: Would anyone let me know the usage of the boost::uniform_int combined with the one argument operator() of the variate_generator? I'd like to change the maximum parameter of the distribution at every step. #include <iostream> #include <boost/random.hpp> int main( void ) { boost::variate_generator< boost::mt19937, boost::uniform_int< std::size_t > > gen( boost::mt19937( 3 ), boost::uniform_int< std::size_t >() ); for ( std::size_t i(1); i < 100; ++i ) std::cout << gen( i ) << std::endl; return 0; } The gcc aborted the compile with the following errors: /usr/include/boost/random/variate_generator.hpp: In member function `typename Distribution::result_type boost::variate_generator<Engine, Distribution>::operator()(T) [with T = unsigned int, Engine = boost::mt19937, Distribution = boost::uniform_int<size_t>]': uniform_int.cpp:9: instantiated from here /usr/include/boost/random/variate_generator.hpp:107: error: no match for call to `(boost::uniform_int<size_t>) ( boost::random::detail::pass_through_engine<boost::random::detail::pass_through_engine<boost::mt19937> >&, unsigned int&)' With regards, Masaki Nomura

2 2

Re: [boost] Re: 1.32 documentation issue
by JOAQUIN LOPEZ MU?Z 06 Nov '04

06 Nov '04

----- Mensaje original ----- De: Beman Dawes <bdawes(a)acm.org> Fecha: Martes, Noviembre 2, 2004 9:06 pm Asunto: Re: [boost] Re: 1.32 documentation issue > At 12:39 PM 11/2/2004, Robert Ramey wrote: > > >Actually after poking around in other pages - specifically the > spirit >documentation I find that its not uncommon for links to > point to > >www.boost.org rather than have a relative path that would keep > the > session > >in the same documentation tree. This would seem to conflict > with the > idea > >that boost should be usable as a CDROM distribution. > > > >I don't know if there's time for this but the inspect should be > enhanced > to > >check for this and all the offending global links should be > altered to > >relative ones. > > We tried to do that at one time, but gave up because it is a bit > tricky. > Some absolute www.boost.org references are correct, such as to the > LICENSE > file, so it isn't as simple as just searching for "www.boost.org/" > in .htm? > files. > > I've just taken a look (using grep) at the CVS RC (not the > tarball) and it > seems that while the problem does happen, particularly in spirit > and graph > libraries, it isn't really horrible. > > I'd say we can wait until after the release ships, then see if we > can come > up with some logic that doesn't generate a lot of false positives. > Automatic detection aside, would it be possible to fix the offending links manually before release? Maybe it's not that straightforward, if those links are generated from BoostBook source. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

3 2

[1.32 release] Runaway compile process in regression tests
by Zak Kipling 06 Nov '04

06 Nov '04

I'm running the regression tests for the 1.32 draft tarballs -- no problems on x86/debian, but on an amd64/gentoo (gcc 3.3.4 prerelease) machine it still hasn't completed after nearly two days, apparently due to a runaway cc1plus process, compiling numeric/conversion/test/converter_test.cpp. This appears to be due to a bug in gcc -- see http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18300. I should have a patch to gcc tested soon; I don't think it's a problem with Boost as such, although there might be a workaround for current compilers (I don't think I understand the numeric converters library well enough to be sure). Zak.

2 1

Compile fix for Boost.Test for aCC on HP-UX
by Ferdinand Prantl 05 Nov '04

05 Nov '04

Hello, you can find a small fix for boost/test/detail/basic_cstring/basic_cstring.hpp attached and a dirty workaround for a compiler bug in aCC for boost/test/unit_test_suite.hpp. As of the latter; the compiler seems to get confused by equal names in different namespaces and classes; renaming Impl helps or making the previous structure friend helps too (this I chose to minimize the change). Ferda HP-UX B.11.11 U 9000/800 aCC: HP ANSI C++ B3910B A.03.37 <<basic_cstring.hpp.diff>>

2 1

[Config] proposal of new macro, BOOST_DECL :follow up for boost::noncopyable & vc7.1 & dll
by Francis ANDRE 05 Nov '04

05 Nov '04

Hi Developers When using the boost.thread package as a dll under vc7.1, one gets for each reference of boost::mutex the C4275 warning for the inherited noncopyable class (in fact 5 times, one for mutex, one for timed_mutex, one for try_mutex, one for thread, and one for thread_group!!) c:\boost\include\boost-1_32\boost\thread\mutex.hpp(34) : warning C4275: non - DLL-interface 'boost::noncopyable_::noncopyable' used as base for DLL-interface 'boost::mutex' In order to avoid such warning, it would be nice to redefine -- again(sic) -- the BOOST_DECL macro and put it as a decorator for all dependent classes used by the boost libraries that could be generated as a dll.(thread, regex, filesystem, and so on). With the empty string as default value in boost/config.hpp, it would the responsability of each boost libraries to setup BOOST_DECL to the correct value when using it as a dll. Cheers Francis ANDRE

5 7