- Boost - lists.boost.org

IOStream mmap too restrictive?
by Neal D. Becker 01 Sep '04

01 Sep '04

"Files must exist before being opened, and are never truncated; attempting to write past the end of a file results in an error. " This is quite restrictive, compared to the capabilities of posix mmap. Are these restrictions really necessary?

2 3

lambda; ->*
by Goran Mitrovic 01 Sep '04

01 Sep '04

Is it normal that boost lambda's member pointer operator fails on member functions that have throw() modifier? Compiler is VC7.1. Is there a solution?

1 0

integer tests not working on 64 bit platforms
by Markus Schöpflin 01 Sep '04

01 Sep '04

Currently the integer tests report an error on 64 bit platforms, as can be seen on http://tinyurl.com/6whtg The following code in integer_test.cpp causes the problem: PRIVATE_FIT_TESTS( int_t, least, long, LONG_MAX ); PRIVATE_FIT_TESTS( int_t, fast, long, LONG_MAX ); PRIVATE_FIT_TESTS( uint_t, least, unsigned long, ULONG_MAX ); PRIVATE_FIT_TESTS( uint_t, fast, unsigned long, ULONG_MAX ); The first line expands to: do { long v = LONG_MAX ; PRIVATE_FIT_TEST(int_t, 32, least, v); v >>= 1; PRIVATE_FIT_TEST(int_t, 31, least, v); v >>= 1; ... } which becomes: do { long v = LONG_MAX ; BOOST_TEST( int_t < 32 > :: least ( v ) == v ); v >>= 1; BOOST_TEST( int_t < 31 > :: least ( v ) == v ); v >>= 1; ... } which of course alway fails on 64 bit platforms because LONG_MAX is a 64 bit number. A possible fix would be to start the test with 64 bit and not 32 on 64 bit platforms. But how can 64 bit platforms be detected? Probably with #if (LONG_MAX > INT_MAX) ... #endif or something like this. Another fix (which ignores 64 bit values alltogether) would be to change the lines above to not use long and LONG_MAX but int and INT_MAX. Opinions, anyone? Markus

3 2

Boost::Python exception translation not working?
by Dan 01 Sep '04

01 Sep '04

I attempted a simple exception translation that was almost an exact copy of the examples in the docs. struct MySQLException : public std::exception { MySQLException( const char * what = "MySQL Exception" ) : std::exception( what ) {} }; And then in a class's default constructor I throw this exception. I tried exposing it as both a function and a functor: void mysql_translate( MySQLException& e ) { PyErr_SetString( PyExc_RuntimeError, e.what() ); } BOOST_PYTHON_MODULE(pymysql) { // regular expression translation py::register_exception_translator< MySQLException >( &mysql_translate ); } And... struct mysql_translate { void operator ()( const MySQLException& e ) const { PyErr_SetString( PyExc_RuntimeError, e.what() ); } }; BOOST_PYTHON_MODULE(pymysql) { // regular expression translation py::register_exception_translator< MySQLException >( mysql_translate() ); } Both had the same result when I created the exported object. "RuntimeError: unidentifiable C++ exception". What am I doing wrong? Thanks, -Dan

2 4

Re: [boost] CodeWarrior and alignment_of (help!)
by JOAQUIN LOPEZ MU?Z 01 Sep '04

01 Sep '04

----- Mensaje original ----- De: Doug Gregor <dgregor(a)cs.indiana.edu> Fecha: Miércoles, Septiembre 1, 2004 6:46 pm Asunto: [boost] CodeWarrior and alignment_of (help!) > Some of the Graph tests are failing on CodeWarrior due to some > strange > interaction within alignment_of, where it doesn't like > mpl::size_t. I > don't have the compiler to test on :( > > Here's one example of the failures: http://tinyurl.com/4dapg > I think current results are built from a CVS merge in progress by Aleksey, and as such one shouldn't pay them too much attention until they're stabilized. The particular problem you refer to has to do (I think) with some haeder making the symbol size_t ambiguous; this pops with variations in other tests. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

tuple and std::set question
by Trey Jackson 01 Sep '04

01 Sep '04

All, I was trying to use the tuple in a set, and the STL complained about not having an operator&() - I think. My question is, is tuple supposed to be usable in a set? I read the STL docs: no mention of operator&() required. I tried both STLport and gcc's stl. I'm using gcc 3.2.2, and the snippet of code is real easy: ,---------------- | #include <boost/tuple/tuple.hpp> | #include <set> | | using namespace std; | | using namespace boost; | | int | main(int, char *[]) | { | typedef tuple<int, int, int> Things; | | std::set<Things> mySet; | | Things oneThing(1,2,3); | mySet.insert(oneThing); | | return 0; | } `---------------- tia, TJ -- Trey Jackson tjackson(a)ichips.intel.com "Nobody can give you wiser advice than yourself." -- Cicero

2 2

Nightly CVS documentation
by Jurko Gospodneti� 01 Sep '04

01 Sep '04

1 0

regression tests broken?
by Bronek Kozicki 01 Sep '04

01 Sep '04

I'm no longer able to run regression tests, and quick look at ftp://fx.meta-comm.com/boost-regression/CVS-HEAD/ reveals that few others have this problem too. Apparently bjam is called from regression.py in wrong way, or Jamfile is broken, as bjam.log ends with: don't know how to make None don't know how to make -sALL_LOCATE_TARGET=C:\MinGW\Bronek\boost\results ...found 2 targets... ...can't find 2 targets... B.

3 5

Re: [boost] IOStreams formal review start (memory-mapped files)
by Jonathan Graehl 01 Sep '04

01 Sep '04

Generally: four stars, two thumbs up, and a partridge in a pear tree. I'm interested in directly using the Windows/POSIX mmap portability wrapper; might we promote it out of the detail namespace? That is, I may want to use mmap to perform some explicit memory swapping/caching. It looks like I could first create a file of the desired size, then acquire a mapped_file_resource - but then I don't get access to the actual pointer to memory (it's hidden in the private _pimpl). This is fine; what I really want is a portability wrapper for mmap :) Thanks, Jonathan

2 1

Problems with Boost.Random docs
by Daryle Walker 01 Sep '04

01 Sep '04

* BOOST_ROOT/libs/random/random-misc.html Mentions "generator_iterator" at the top, but there's nothing else about it. Does it have to be a secret that "random_number_generator" can now be accessed in its own header ("random_number_generator.hpp")? The web page still places it in "random.hpp" (which does include the previous header). I just noticed that "random.hpp" doesn't mention anything about a generator-iterator either. * BOOST_ROOT/libs/random/random-distributions.html Mentions that users that want an iterator interface should just put an adaptor over the existing generator interface. That sentence gives a link to the Iterator Adaptor library, instead of its own generator-iterator (which I guess is good in a way since it's MIA). There's a problem with this arrangement: * BOOST_ROOT/libs/iterator/doc/index.html Of its list of specialized adaptors, there's is NO generator adaptor! (There is an adaptor that goes the other way, i.e. output iterator to unary function, but we need generator to input iterator.) The sum up, it looks like the Random guys dumped their private generator adaptor for iterators, expecting the Iterator guys to replace it, but the latter guys never did. Unless the generator adaptor is hidden somewhere else, can someone add it (back). BTW, a generator would have to support a Boolean conversion where TRUE means it's still usable and FALSE means it isn't. (This is a new semantic for generators.) This would allow a reasonable way to implement the end-input-iterator. -- Daryle Walker Mac, Internet, and Video Game Junkie darylew AT hotmail DOT com

1 0