- Boost - lists.preview.boost.org

Re: [boost] Re: static sized strings
by Reece Dunn 10 Sep '04

10 Sep '04

Rob Stewart wrote: >From: "Reece Dunn" <msclrhd(a)hotmail.com> > > [1] How do I pass the buffer and capacity to the implementation while > > allowing both fixed_string_base and fixed_string access to the >basic_string > > interface? I have chosen to route the data through one of the >basic_string > > constructors, requiring a const_cast. This is the best approach AFAICS. > >I haven't looked at your recent implementation, and I don't >presently have the time, so I really don't know what you're >talking about. If you can explain the dependencies to which you >refer, I'd appreciate it. I think it has bearing on the swap() >discussion. I have a class called detail::fixed_string_impl that has three data members: size, capacity and a pointer to the data buffer (len, cap and str respectively). Doing this allows the virtual functions to be removed and the implementation moved out of fixed_string. The fixed_string class can then pass its buffer and the capacity template parameter to the implementation via a constructor. The problem is that fixed_string is derived from fixed_string_base, which itself is derived from detail::basic_string_impl. This is needed to allow both fixed_string and fixed_string_base to have access to the full basic_string interface (like was done in the old implementation). The solution I have taken is to use the basic_string_impl( const CharT * s, size_type l ) constructor and use this to pass the buffer (s) and capacity (l) to the string. This is not ideal since it requires a const_cast to extract the buffer. This required modifying the constructor implementation in basic_string_impl to forward the parameters to Base and not implement them using append. The old implementation had the same problem with swap as the current one, since the string would get clipped when it is copied. For now, I will leave the implementation of swap as is (i.e. clipping and not signalling an exception), but add a note on this issue in the docs so that users are aware of it. Another problem I can forsee is using a custom char_traits or format_policy class, e.g. boost::fixed_string< 20, char, std::char_traits< char > > s1; strcpy( s1, "Hello" ); // ok boost::fixed_string< 20, char, win_char_traits< char > > s2; strcpy( s2, "There" ); // error This is the problem shared with basic_string when implementing functions using std::string. I have also fixed the problems relating to using the strxxx functions with Borland: the compiler may set them as instrinsic and add a macros defining strxxx as __strxxx__. To get around this, I now detect if they are macros and the Borland compiler is being used, then create an inlined function to the __strxxx__ version. Regards, Reece _________________________________________________________________ Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo

2 1

filesystem exists/is_directory suggestion
by Martin 10 Sep '04

10 Sep '04

The filesystem status operations exists & is_directory will both throw if the status of the file can't be determined (access problems or race conditions). I think this makes the operations overly complicated to use since you need to enclose them in try catch blocks. You normally guard operations that use files in some way but it is easy to forget try/catch around exists / is_directory. (Noticed that exists has been updated in CVS to return true on error which is probably ok but I think is_directory need to be updated as well.) I think a good solution would be to have a non-throwing overload where the return value in case of error is specified. bool is_directory(const tdFS::path& aPath, bool aErrorRet) { try { return tdFS::is_directory(aPath); } catch (...) { return aErrorRet; } } Another way could be to have a "is_accessable" operation that determines if the status of the entry can be determined but it will not solve the racing problem. usage: for (directory_iterator itr(mypath); itr != directory_iterator(); ++itr) if (is_directory(*itr, true)) continue; // skip directories or for (directory_iterator itr(mypath); itr != directory_iterator(); ++itr) if (!is_accessable(*itr) || is_directory(*itr)) continue; instead of for (directory_iterator itr(mypath); itr != directory_iterator(); ++itr) try { if (is_directory(*itr)) continue; } catch (...) { continue; }

4 8

[admin] Overlapping reviews -- should this be allowed?
by Jeff Garland 10 Sep '04

10 Sep '04

Hi all - Our review wizard, Tom B. has suggested to me that we allow subsequent reviews to begin even if a previous review has been extended. This means 2 or more reviews will possibly be running in parallel. I don't believe we have a written policy, but I'm quite certain we have never run reviews in parallel. I also think it is a very bad idea. I believe reviews will suffer because people that want to review all the libraries in progress and participate in the discussion will not have enough time to do so. For most of us boost is a part-time thing -- we only have so much time per week to participate... This leads me to the conclusion that the current review schedule needs to be redone. 1st because IOStreams has been extended and 2nd because we agree that the current schedule is too aggressive. See http://lists.boost.org/MailArchives/boost/msg11034.php I realize it is a major pain to rearrange the agreed dates, but I don't think overlapping reviews is the answer. Thoughts? Jeff

11 26

Filesytem compile problem
by Martin 10 Sep '04

10 Sep '04

I get a compile error in boost/filesystem/operations.hpp when building an application that also uses the date_time library (using current CVS files). The problem is that date_time includes boost/operator.hpp which defines the template boost::equivalent. The name collides with boost::filesystem::equivalent. Reordering the include files (filesystem before date_time) fixes the problem for me but since boost/operators.hpp is used in a lot of libraries it might not always be that easy.

2 1

intel problems
by Pavol Droba 10 Sep '04

10 Sep '04

Hi, Most of tests in intel-win32-7.1-vc6-stlport-4.5.3 toolset are failing for the string_algo lib. Here is the output: http://tinyurl.com/5l2k4 All are linkinig errors. So there is probably an error in my jamfiles. Can somebody give me a hint what's wrong? Thanks, Pavol

1 0

RE: [boost] [filesystem] large file support test added
by Ben Hutchings 10 Sep '04

10 Sep '04

Beman Dawes <bdawes(a)acm.org> wrote: <snip> > Normally we would have a regression test that identifies the systems > which fail. But we don't test large file support directly because to > do so would be a burden on those who run the regression tests; such > tests would chew up gigabytes of disk space and might also be very > slow. <snip> why so? Unix file-systems support sparse files, as does NTFS.

2 1

[serialization] - xml bug?
by Jeff Flinn 10 Sep '04

10 Sep '04

Robert, The following demonstrates a problem with serialization of strings conataining '<' and/or '>'. This is on VC7.1 using serialization20 with boost_1_13_0. I'm not in a position to check this on the CVS version at this time. Is this still an issue? Thanks, Jeff ///////////////////////////////////////// #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> #include <string> #include <fstream> int main(int argc, char* argv[]) { std::string lStringOut("<empty>"); { std::ofstream lOut( "C:/xmltest.xml" , std::ios_base::out | std::ios_base::trunc ); if( lOut.good() ) { boost::archive::xml_oarchive oa( lOut ); oa << boost::serialization::make_nvp( "MyString", lStringOut ); } } std::string lStringIn; { std::ifstream lIn( "C:/xmltest.xml" ); if( lIn.good() ) { boost::archive::xml_iarchive ia( lIn ); ia >> boost::serialization::make_nvp( "MyString", lStringIn ); } } assert( lStringOut == lStringIn ); return 0; }

1 0

[filesystem] large file support test added
by Beman Dawes 10 Sep '04

10 Sep '04

Something special (defining some preprocessor macro) has to be done for 32-bit POSIX-like operating systems to enable support for file sizes greater than can be represented by 31 bits. The exact preprocessor macro name that has to be defined varies from system to system. Boost Filesystem currently supplies the correct macro for Linux, but not some of the UNIX systems. Normally we would have a regression test that identifies the systems which fail. But we don't test large file support directly because to do so would be a burden on those who run the regression tests; such tests would chew up gigabytes of disk space and might also be very slow. It recently occurred to me that sizeof(stat.st_size) > 4 is a pretty good surrogate for an actual test on a large file. While in theory it could yield a false positive, in practice it should identify whether or not large file support has been correctly enabled. It should never yield a false negative. Accordingly, I've added a new regression test, large_file_support_test.cpp. Initially it should pass on Windows and Linux systems, but fail on any UNIX systems that need additional macros defined. We can then experiment with adding additional macro definitions to get the test to pass on any systems it is currently failing. --Beman

1 0

dynamically created program options
by Rodolfo Schulz de Lima 10 Sep '04

10 Sep '04

Hi, I wonder if it's possible to handle dynamically created program options with boost::program_options. For instance: I have a config file that specifies configure options for dynamic loaded modules in my project. So the cfg file would be: [module.foo1] module_file = foo.so listen_port = 564 enabled = true [module.foo2] module_file = foo.so listen_port = 565 enabled = false [module.bar1] module_file = bar.so listen_port = 145 enabled = true As can be seen, foo1, foo2 and bar1 names are defined by the user in the configure file, and the program doesn't know about them before parsing the configuration. But if we parse that (somehow), an exception saying that the option "module.foo1.module_file" is invalid will be thrown. Can I manage the options_description "on the fly"? Another less complicated situation would be something like this: modules = foo1 foo2 bar1 [module.foo1] . . . [module.foo2] . . . [module.bar1] . . . So when parsing the "modules" option, we would know the new options name and it would be a matter of appending more option descriptions to the object. Regards, rod

2 1

ublas::vector from std::vector
by Neal D. Becker 10 Sep '04

10 Sep '04

I noticed there is no constructor for ublas::vector that takes a pair of iterators. Is there no direct way to construct a ublas::vector from a std::vector? Could there either be a constructor: (STL-style) template<typename input_iterator> vector (input_iterator begin, input_iterator end); or: (range-style) template<typename cont_t> vector (cont_t const&)

1 0