April 2003 - Boost-users - lists.preview.boost.org

Re: [boost] lexical_cast problem
by Terje Slettebø 18 Apr '03

18 Apr '03

>From: "Terje Slettebø" <tslettebo(a)chello.no> > Hm, just now, I sent a reply to a similar problem report at Boost-Users > list, about why whitespace skipping is turned off in the new lexical_cast > version. It was to avoid having things like "lexical_cast<std::string>(' '); > // Space character" work, which used to throw an exception. I meant it was to avoid having it throw an exception, in other words, to _have_ it work. Terje

1 0

Re: [boost] lexical_cast problem
by Terje Slettebø 18 Apr '03

18 Apr '03

>From: "Vladimir Prus" <ghost(a)cs.msu.su> >Subject: [boost] lexical_cast problem Not again. :) > I've just updated my project to use boost 1.30 and also gcc 3.2 and see > exception which reads: > > bad lexical cast: source type value could not be interpreted as target, > Target=i, Source=Ss Ah, yes, good ol' g++ with its wonderfully descriptive type_info names... ;) > The simple reproduction recipe is: > > #include <boost/lexical_cast.hpp> > #include <string> > > int main() > { > boost::lexical_cast<int>(std::string(" 0")); > } "Houston, we have a problem..." Hm, just now, I sent a reply to a similar problem report at Boost-Users list, about why whitespace skipping is turned off in the new lexical_cast version. It was to avoid having things like "lexical_cast<std::string>(' '); // Space character" work, which used to throw an exception. However, it seems this change has had undesired side-effects, in that code like the above now fails to work. Frankly, I don't know what to do with this... > Note the space in string. The exception does not occur with gcc 2.95, but only > with 3.2. I wonder if this should/can be fixed? Personally, I'd expect this > code to work. > > And a related note: the exception's what() above is not quite helpfull. The current version in the CVS has got a changed exception handling, where bad_lexical_cast now looks like this: class bad_lexical_cast : public std::bad_cast { public: // ... const std::type_info &source_type() const const std::type_info &target_type() const virtual const char *what() const throw() { return "bad lexical cast: " "source type value could not be interpreted as target"; } }; In other words, what() now returns the original text, and the type_info objects for the types used is now returned by member functions. > Is it > possible to output the stringstream content (if source -> stringstream > conversion was successfull?) Yes, that's possible. I guess I'll leave it to Kevlin to comment on this. Thanks for the report, to both you and Tom Matelich at Boost-Users list. Regards, Terje

1 0

lexical_cast regression
by Tom Matelich 18 Apr '03

18 Apr '03

I haven't gone and looked at the boost ml archives to track the reason for this down, but the addition of stream.unsetf(std::ios::skipws) in lexical_cast caused some of our code to fail. I've commented it out locally, and was wondering why it was decided this is more appropriate default behavior. Thanks, Tom ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

2 1

kinda, sorta, like an iterator for dynamic_bitset, but not really.
by Andy Klapper 18 Apr '03

18 Apr '03

I'm currently writing some code for a client where I make extensive use of dynamic_bitsets to determine very quickly which options are appropriate based on a number of criteria. I set up before hand dynamic_bitset for each person, place, and thing which are then bitwise AND'd together to determine where the points of intersection are. What I need is something with an iterator syntax that returns the indexes whose values are set. I could of course just loop through looking for set values, but that seems silly when you think that entire blocks of bits can be skipped if the block has a value of zero. So my questions is, before I set out to write some code, has somebody already done this? It seems like a logical use of dynamic_bit set so it does not seem implausible. If not, any suggestions or warnings? Thanks, Andy

2 1

Understanding MPL
by Edward Diener 18 Apr '03

18 Apr '03

I am trying to understand how the MPL library works. I first read "The Boost C++ Metaprogramming Library" by Aleksey Gurtovoy and David Abrahams. When I get to Basic Usage I see this: template< typename Condition , typename T1 , typename T2 > struct if_ { typedef /*unspecified*/ type; }; I am really having trouble understanding the next to last line. How can "typedef type;" be a valid C++ construct and, if it is, what does it mean ? I always though that a typedef must specify an actual type but I don't see any in the expression above.

7 27

RE: [Boost-Users] Re: command line parsing
by Ben Hutchings 17 Apr '03

17 Apr '03

E. Scott Larsen wrote: > Just FYI to this thread in general. It's been a while since I've messed > with WinMain and command lines, but if my memory serves me right (it was > at a different place, so I don't have the code here), there's some > additional whacky issues here. WinMain recieves the first arguments > differently depending on how it is run: > *) from MSVC debugger > *) from a commandline > *) by clicking on an shortcut icon > *) by clicking in an explorer window I find this hard to believe, as WinMain is just given a pointer into the command line returned by GetCommandLine(), incremented past the program name and any following space. > don't remember which is which, but one gives just the executable name, > one gives the full path to the executable (seems like there's two > variations on this, like one is quoted and one is not, don't remember > the details), and it seems like there's one that doesn't give the > executable at all. I can imagine there's some variation in program name; it's normal for programs to be started with short names rather than their full filenames. The quoting is weird though; that should be removed before you see it. > Also, I _distinctly_ remember something very whacky when I used unicode, > I don't remember the details, but the commandline arguments were a > significant pain for a program built with unicode support using the > unicode windows entry function. That still tastes bad in my mouth. According to the docs, this ought to work: wchar_t ** argv; int argc; argv = CommandLineToArgvW(GetCommandLineW(), &argc); but I've never tried it; maybe it doesn't. If you were using WinCE then I feel your pain and hope you've recovered.

3 2

RE: [Boost-Users] Powers
by Jim.Hyslop 17 Apr '03

17 Apr '03

basicxml2011 [mailto:basicxml2011@yahoo.com] wrote: > Is their any library or function that can raise a number to a power? Do you mean besides the standard pow() function? -- Jim

1 0

mpl::is_same<Class<A>, T> ?
by Markus Werle 17 Apr '03

17 Apr '03

Hi! Is there a convenient way inside boost::mpl to check whether a certain type T is a template Class<A> with A unspecified? Markus -- Compile time analytic differentiation? Yes, at http://daixtrose.sourceforge.net/

2 1

How to build regex on an Alpha-Dec (True64)?
by Ralf Hupfer 17 Apr '03

17 Apr '03

Hi All, I am having trouble with regex on an Alpha-Dec with Compaq's compiler 6.3. I've got regex to compile with generic.mak as follows: CXX="cxx" CXXFLAGS="-inline-none" -D__USE_STD_IOSTREAM -DBOOST_REGEX_NO_EXTERNAL_TEMPLATES" LINKER="cxx" gnumake generic.mak but then it comes to the link step an error occures: ld: Unresolved: main Nevertheless, boost_regex.so is being build, but when I try to link an application with boost_regex.so I do get loads of errors concerning the C++ standard library: ld (prelink): /usr/lib/cmplrs/cxx/V6.3-008/libcxxstd.a(cxxl_std_init.o): std::cout: multiply defined Warning: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::npos defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT Warning: std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> >::npos defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT Warning: std::locale::collate defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT Warning: std::locale::ctype defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT Warning: std::locale::monetary defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT Warning: std::locale::numeric defined as GLOBAL RCONST but is defined in a shared lib as a GLOBAL OBJECT ... Does anybody know what's wrong there? Both -D__USE_STD_IOSTREAM and -DBOOST_REGEX_NO_EXTERNAL_TEMPLATES are necessary to compile regex at all. -D__USE_STD_IOSTREAM tells the compiler to use the standard iostream library instead of the older non-standard version (which is the default for this platform). Without -DBOOST_REGEX_NO_EXTERNAL_TEMPLATES the compiler complains about some templates. I also tried to compile regex with gcc, but again I could not link the application. Linking gcc-compiled (c++) libraries with Compaq's object files and vice versa does not seem to work in generall. Building everything with gcc failed due to seome other incompatibilities. Thanks for any help. Ralf

4 3

RE: [Boost-Users] Re: command line parsing
by Ben Hutchings 17 Apr '03

17 Apr '03

I wrote: > That is correct. Windows programs receive given a single string of > arguments to parse whereas Unix programs receive a vector which can > be passed to main unchanged by the startup code. ... > A full solution will need to be a little bit more complicated. There > isn't a specification of how Windows command-lines should be generated > or interpreted from an array of strings... What I should have written is: Windows programs receive a single string of arguments to parse whereas Unix programs receive an array which can be passed to main unchanged by the startup code. ... A full solution will need to be a little bit more complicated. There isn't a specification of how Windows command-lines should be converted to or from an array of strings... By the way, the startup code or other command-line parser does not have to handle redirection or variable substitution; this *is* done by the command processor (which implies that it does parse the command- line, even though it doesn't give you the results of this).

4 4