- Boost-users - lists.boost.org

compiling 64 bit boost on Mac OS 10.5
by Tom Smith 04 Mar '08

04 Mar '08

Hi, I've been trying and failing to get a 64 bit build of the boost libraries on 10.5. I always seem to end up with 32 bit binaries. I've tried various permutations involving bjam address-model=64 (I last tried to build a week ago so I'm hazy on the actual command line). I only need x64, not ppc64. Cheers, Tom

3 5

iterator_adapter/iterator_tag/std::distance bug
by Martin Wartens 04 Mar '08

04 Mar '08

I used boost::iterator_adapter with a std::vector iterator, all parameters set to boost::use_default. (So this is the same as a plain std::vector iterator) With this configuration, the iterator tag is recognized as "input_iterator_tag" not "random_iterator_tag" as it must be. This hurts because the iterator tag is used to select the right std::distance function. I could solve the problem by explicitly setting the traversal type template parameter to "std::random_access_iterator_tag", while "boost::random_access_traversal_tag" also didn't work. The compiler is VC8.

2 2

serialization - xml
by Ambardekar, Mukund 03 Mar '08

03 Mar '08

Hi, I used boost serialization to write my data in binary format as in below way - std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::binary_oarchive oa(ofs); //write oa & m_myclass; // object of CMyClass type containing some data (say name) This worked correctly. Now I wanted to write into xml format. Did the below changes std::ofstream ofs(catalogFileName.c_str(),std::ios::binary); boost::archive::xml_oarchive oa(ofs); //write oa & m_myclass; But it is giving me below error - d:\tools\boost-1feb08\boost_1_34_1\boost\archive\basic_xml_oarchive.hpp( 83) : error C2027: use of undefined type 'boost::STATIC_ASSERTION_FAILURE<x>' Can you help me to identify the problem? FYI - I have made no changed CMyClass looks like as below. And I have made no changes in it (same one used for binary and xml) class CMyClass { public: Part(const std::wstring& Name); private: std::wstring m_name; friend class boost::serialization::access; template<class Archive> void serialize(Archive& ar, const unsigned int) { ar & m_name; } }; Thanks in advance. Regards, Mukund

2 1

Minimal human-readable boost::serialization ?
by Eric Newhuis 03 Mar '08

03 Mar '08

I looks like the last time this was mentioned in this mailing list was in 2005. Has anyone considered anything like boost::archive::json_oarchive etc.? JSON looks easy to read in log files. I want name-value pairs in my serialization. But I don't want the verbosity of XML. I think JSON would be ideal.

2 1

Better debugger support for boost::optional with VS2005 compiler
by Daniel Krügler 03 Mar '08

03 Mar '08

Hello, just recently I introduced boost::optional as default type to represent optional/nullable types in a sub-project of our company, which generates C++ code out of UML models. While this worked seamless, programmers not yet used to boost::optional complained that they could not really see, what such type "contains" if they try to observe it in the debugger. I searched for a solution, only finding the discussion around the article "how to see the actual value of a 'boost::optional<__int64>' in VS2005 debugger" from 2006-07-26, which did not gave a satisfactory answer. Please ignore this message, if the here described solution (or a better one), which I worked out this weekend, is already well-known (but I would appreciate if someone could give me a link to the already known solution). My approach was to modify the unofficially supported debugger configuration file autoexp.dat, which typically is found at ${SomePath}\Microsoft Visual Studio 8\Common7\Packages\Debugger\autoexp.dat If you never have heard of this file, you should make a copy of it, before doing the modifications I discuss here. A more detailled description what can be done with this file, can be read here (but you don't need to do that, if you just want to get the results I propose here): http://www.virtualdub.org/blog/pivot/entry.php?id=120 Just use your preferred text editor to open the autoexp.dat, jump at the end of it and go five rows back. You should be +/- 1 row about the line with the content: ; This section lets you define your own errors for the HRESULT display. Go one row back (row 1957 in my file), which should be the end of the so-called [Visualizer] section. Now insert here the text provided at the end of this article and just start you debugger session (you don't need to restart the studio) - you should have a nice view on each boost::optional, even those which are optionals of references. I tested the visualizer for the following boost versions: 1.30.0 1.31.0 1.32.0 1.33.0 1.33.1 1.34.1 and all of them worked. I would like to thank the boost developers for their great contributions (where boost::optional is only one of them) and I hope you can even more enjoy boost::optional with this trick. Unfortunately I have not found similar solutions for other compilers yet, e.g for gcc, which probably is due to my lack of experience with them but I would appreciate if someone else publishes them, too. Greetings from Bremen, Daniel Krügler ;------------------------------------------------------------------------------ ; Visualizer for boost::optional, boost::optional_detail::optional_base ;------------------------------------------------------------------------------ boost::optional<*>|boost::optional_detail::optional_base<*> { preview ( #if($e.m_initialized) ( *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) #else ( "?" ) ) children ( #if($e.m_initialized) ( #(value: *(($T1 *) (&$e.m_storage.dummy_.data[0])) ) ) #else ( "?" ) ) } ;------------------------------------------------------------------------------

3 3

[filesystem] bug???
by Ovanes Markarian 03 Mar '08

03 Mar '08

Hello *! A short question: removing the leaf from the path using the remove_leaf() results in my case (MS VC7.1, boost 1.34.1) in the path with the trailing slash at the end. Next call to remove_leaf() results in removing this trailing slash, which is not correct... Is it a bug, if the doc states to remove_leaf: basic_path& remove_leaf(); Effects: If has_branch_path() then remove the last filename from the stored path. If that leaves the stored path with one or more trailing slash elements not representing root-directory, remove them. Many thanks, Ovanes

1 0

A possible bug in Boost.property_tree with UNICODE XML file?
by Tan, Tom (Shanghai) 03 Mar '08

03 Mar '08

I am trying read_xml() on an xml-format config file, which starts like this: <?xml version="1.0" encoding="UTF-16"> ..... This file is well-formed and verified by "XML Copy Writer", it can also be open by IE or Filefox with no problem. I am using this code snippet to parse this file: try{ read_xml<wptree>(xmlfile,pt); } catch(xml_parser-error e) { e.what(); } and it throws an exception saying "expected <". By looking into the code I found that, this caused by the fact that the xmlfile starts with a Unicode Byte Order Mark (BOM)-0xFFFE . "fffe 3c00 3f00 7800 ......" However, read_xml() cannot handle this BOM byte and throws. It's easy to fix it by simply discard the BOM byte, but I wonder whether there is already a flag to enable/disable this that I am unaware of. Can anyone help to clarify? Thanks, tom

1 0

Re: [Boost-users] templates: parts of functions
by Lang Stefan 03 Mar '08

03 Mar '08

Hi, As I am not overly familiar with boost or it's mpl library yet, I cannot provide any specific advice along these lines. According to your description, my advice would be to create helper template classes or functions that rely only on some of the template parameters of your actual template class, but not N. For these helper functions and classes you do not need partial specialization, and thus no duplicate code. Depending on the complexity of your functions you might also extract those parts that do not depend on N into more template functions. If these helpers need access to private members of your makeTree class, you can declare them friend. (you might need to replicate those friend declarations in your partial specializations) For those parts of the code that do depend on N, but do not need a different implementation, you may still provide templated helper classes. The compiler will still generate multiple class implementations for different values of N, but you won't need to replicate the actual code yourself. Hope this helps. Cheers, Stefan > ------------------------------ > > Message: 6 > Date: Sat, 1 Mar 2008 14:38:34 +0100 > From: "Hicham Mouline" <hicham(a)mouline.org> > Subject: [Boost-users] templates: parts of functions > To: <boost-users(a)lists.boost.org> > Message-ID: <000001c87ba1$8cfe0d30$a6fa2790$@org> > Content-Type: text/plain; charset="us-ascii" > > Hi, > > > > I have a function makeTree whose behavior depends on 3 > different policies > > typename T1 > > typename T2 > > int N > > > > For performance reasons, I need to write the 2- and 3- specializations > alongside the general N case. That is, the 2- and 3- default > > instantiations generated from the N case by the compiler are > not optimal. > > > > As it is impossible to write function templates partial > specializations, I > encapsulate the > > > > template< typename T1, typename T2, int N> void > makeTree(Head<N>&, const > T2&, const C&) ; > > > > function in a class template > > > > template< typename T1, typename T2, int N> class makeTree{ > > void makeTree(Head<N>&, const T2&, const C&); > > }; > > and write the partial specializations <T1,T2,2> and <T1,T2,3>. > > > > > > Now, this makeTree function is quite large. It includes code that : > > . doesn't depend on any of the template arguments > > . code that depends only T1 > > . code that depends only on T2 > > . code that depends only on N > > . code that depends only on <T1,N>. > > > > I am worried about code bloat, and the effect on overall > performance of the > actual code generated even if it is not run > > (I noticed substantial effects depending on the size of > object files and > final shared library generated). > > > > Besides this, I suppose it is a better design to have these > segments of code > put in template functions > > with the right template arguments. > > > > Is it MPL the library that I should look into for such work? > > Perhaps nothing in Boost particularly facilitates this work, > and std c++ > templates are enough, > > > > Thoughts, comments are appreciated, > > > > Rds, >

1 0

Why does runtime-debugging set _STLP_USE_DYNAMIC_LIB?
by Eric Woodruff 03 Mar '08

03 Mar '08

I've been trying to understand why the debug version of boost was attempting linkage with the DLL version of stlport (which I didn't build and I'm not using) and now I see that the runtime-debugging flag is specifying that the dynamic library must be used. This causes the static stlport to use the stlpdx cross namespace rather than the stlpd namespace. It is not obvious to me why debugging should require a dynamic library. Under code inspection this looks like a copy-paste oversight. if [ $(rproperties).get <runtime-debugging> ] = "on" { usage-requirements += <define>_STLP_DEBUG=1 <define>_STLP_DEBUG_UNINITIALIZED=1 ; } if [ $(rproperties).get <*runtime-debugging*> ] = "on" { usage-requirements += <define>_STLP_USE_DYNAMIC_LIB=1 ; } Why would it be checking for runtime-debugging twice? I think the second if should check for <runtime-link> which has the possible values "static" or "dynamic". This problem looks related to this thread: http://lists.boost.org/boost-users/2007/09/30905.php, but I couldn't find the Boost.Build message that is a follow-up to that one. I haven't found a posted solution either. Looking at the CVS HEAD revision of stlport.jam, it hasn't changed since 1_34_1. So if this is a bug, it has not been fixed yet. Thanks, Eric

2 1

Re: [Boost-users] fatal error C1128: object file format limit exceeded : more than 65, 279 sections -------PROBLEM SOLVED
by jono 03 Mar '08

03 Mar '08

i thought i had broken the translation unit in two, but it was a clever illusion ... doh.

1 0