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

[bjam][1.41.0] Build with variant=release, debug fails
by Richard Hadsell 06 Jan '10

06 Jan '10

I am trying to build the Boost 1.41.0 libraries with this command: bin/bjam --user-config=user-config.LINUX_AMDF7 threading=multi instruction-set=opteron address-model=64 variant=release,debug,profile link=static,shared --stagedir=stageF7 stage The last time I built Boost libraries was version 1.35.0. I eliminated all of the other options to narrow the problem to just this option: variant=release,debug The build starts up successfully, if I select just one option for variant (debug or release), but it fails when I try to build both. These are the error messages: error: Duplicate name of actual target: <pstageF7/lib>libboost_date_time.a error: previous virtual target { common%common.copy-libboost_date_time.a.STATIC_LIB { gcc%gcc.archive-libboost_date_time.a.STATIC_LIB { gcc%gcc.compile.c++-gregorian/greg_month.o.OBJ { gregorian/greg_month.cpp.CPP } } { gcc%gcc.compile.c++-gregorian/greg_weekday.o.OBJ { gregorian/greg_weekday.cpp.CPP } } { gcc%gcc.compile.c++-gregorian/date_generators.o.OBJ { gregorian/date_generators.cpp.CPP } } } } error: created from ./stage-proper error: another virtual target { common%common.copy-libboost_date_time.a.STATIC_LIB { gcc%gcc.archive-libboost_date_time.a.STATIC_LIB { gcc%gcc.compile.c++-gregorian/greg_month.o.OBJ { gregorian/greg_month.cpp.CPP } } { gcc%gcc.compile.c++-gregorian/greg_weekday.o.OBJ { gregorian/greg_weekday.cpp.CPP } } { gcc%gcc.compile.c++-gregorian/date_generators.o.OBJ { gregorian/date_generators.cpp.CPP } } } } error: created from ./stage-proper error: added properties: <debug-symbols>on <inlining>off <optimization>off <runtime-debugging>on <variant>debug error: removed properties: <debug-symbols>off <define>NDEBUG <inlining>full <optimization>speed <runtime-debugging>off <variant>release /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/tools/build/v2/build/virtual-target.jam:490: in actualize-no-scanner from module object(file-target)@4821 /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/tools/build/v2/build/virtual-target.jam:135: in object(file-target)(a)4821.actualize from module object(file-target)@4821 /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/tools/build/v2/build-system.jam:717: in load from module build-system /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/tools/build/v2/kernel/modules.jam:283: in import from module modules /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/tools/build/v2/kernel/bootstrap.jam:142: in boost-build from module /netDISKS/cgi/LINUX_AMD64/cgi/vendor/library/boost/boost_1_41_0/boost-build.jam:17: in module scope from module The user-config contains just these 2 lines: using gcc : 4.1.2 : /usr/bin/g++ : <compileflags>"-fPIC -fno-strict-aliasing -fno-tree-ccp" <linkflags>"-fPIC -fno-strict-aliasing -fno-tree-ccp" ; using python : 2.5 : /usr ; It seems possible that the libraries would build correctly in separate steps, but I am concerned that the debug, release, and profile versions will be mixed up. Is this just a problem for the date_time library, or is there something wrong with the build rules? If someone can give me a reliable workaround, I would appreciate it. (I am building with g++ 4.1.2 on a Fedora 7 x86_64 Linux system.) -- Dick Hadsell 203-992-6320 Fax: 203-992-6001 Reply-to: hadsell(a)blueskystudios.com Blue Sky Studios http://www.blueskystudios.com 1 American Lane, Greenwich, CT 06831-2560

2 3

Reducing the size of .obj file which does serialization
by Gokulakannan Somasundaram 06 Jan '10

06 Jan '10

Hi, I have a parent class which has around 200 sub-classes. While serializing the entire hierarchy, i have lot of base class pointers and STLs of base class pointers. From the documentation, i found that if i have to serialize the polymorphic pointers, then i have to do a BOOST_EXPORT of all the relevant subclasses. But after doing that, i find my .obj file to be around 37MB. I am using the text archive in my code. When i tried the polymorphic_text_archive, the size only reduced by 3MB. Can someone help me on how i should go about reducing the .obj file size? Some of my questions are a) One approach, i could think of is to create a virtual function in the hierarchy, which just takes the Archive and does the serialization. This would help me, because i don't need to have the BOOST_EXPORT, which increases the size of the .obj file. b) Why is BOOST_EXPORT required for polymorphic_text_archive? Since it anyway uses virtual functions instead of templates, why does it need the derived class definition. Thanks, Gokul.

2 5

[Filesystem] directory iteration throws
by gast128 06 Jan '10

06 Jan '10

Hello all, I was writing a simple application to enumerate files sorted on their size. I used the recursive_directory_iterator for this task. However on my Windows XP machine the directory iteration stops because of a throwing constructor. It is easy reproducable, by creating a directory with contents and remove all security attributes (except for the system account, leave there all permissions enabled) and use that directory for iteration, i.e. my c:\temp\bla directory contained 1 file and I removed the security attributes for the directory Bla. Tracing thru the source code of Boost.FileSystem the directory is opened with a '::FindFirstFileA' call which returns an 'INVALID_HANDLE_VALUE', resulting in a throw. This stops further iteration of the directory. Is there a way to work around this?

1 0

Return value of bool boost::interprocess::shared_memory_object::remove(const char * name)
by Jarno 06 Jan '10

06 Jan '10

Hello, It seems to me that the static function bool remove(const char * name) in class boost::interprocess::shared_memory_object returns false when it should return true and vice versa. The documentation on boost interprocess says: "The remove operation might fail returning false if the shared memory does not exist, the file is open or the file is still memory mapped by other processes" However, the following simple test program for example prints "should not print" on Windows XP 64. I'm compiling with MSVC 2008 and using boost 1.41.0. /* Test program */ #include <boost/interprocess/shared_memory_object.hpp> #include <iostream> int main(int argc, char* argv[]) { using namespace boost::interprocess; using namespace std; { shared_memory_object shm(open_or_create, "foo", read_write); } if(shared_memory_object::remove("foo")) cout << "should print" << endl; if(shared_memory_object::remove("foo")) cout << "should not print" << endl; return 0; } /* Test program ends */ Am I interpreting the documentation correctly? Regards, Jarno

2 1

Thread is not interrupted
by Lloyd 06 Jan '10

06 Jan '10

Hi, I am using boost::asio to create a server. When a new connection is established (inside the asio read handle) a new thread (boost::thread) is created. The thread object pointer is kept in a static vector. When the next connection is established, the thread pointer available in the static vector is read and interrupted. But the thread is not interrupted! The thread is just an infinite loop, in which interruption is enabled. I have wrttien the exception handler also. Not able to understand why this is not working in this situation. Otherwise this works. To make this clear given below are the code snippets... //ASIO based read handler void TCPConnection::HandleRead(const boost::system::error_code& Error) { TCPConnection::Threadvector[0]->interrupt(); //Threadvector is static variable in the class, for easy reading written like this boost::thread *th=new boost::thread(boost::bind(&TCPConnection::ExecutionRoutine,shared_from_this(),Command)); TCPConnection::Threadvector.push_back(th); } //ExecutionRoutine looks like this int TCPConnection::ExecutionRoutine(string Command) { try { while(1) { boost::this_thread::interruption_point(); } } catch(boost::thread_interrupted const&) { return 0; } return 0; } Please share some idea... Thanks, Lloyd ___________________ Happy New Year 2010

1 0

[GIL] Packing to 6 bytes
by Eloi Du Bois 06 Jan '10

06 Jan '10

Hi, I have a question: I need to create a view that hold 12 bit channel data packed as follow: 0000R0000G0000B0000;0000R0000G0000B0000;0000R0000G0000B0000;.... Meaning that each channel is shifted by 4 bits. The pixel is holden on 48 bits = 6 bytes. I have created my view like as follow: /// My 12 bits special rgb packed to 6 bytes typedef const packed_channel_reference<uint64_t, 04, 12, true> rgb12_packed_channel0_t; typedef const packed_channel_reference<uint64_t, 20, 12, true> rgb12_packed_channel1_t; typedef const packed_channel_reference<uint64_t, 36, 12, true> rgb12_packed_channel2_t; typedef mpl::vector3<rgb12_packed_channel0_t, rgb12_packed_channel1_t, rgb12_packed_channel2_t> rgb12_packed_channels_t; typedef packed_pixel<packed_channel_value<48>, rgb12_packed_channels_t, rgb_layout_t> rgb12_packed_pixel_t; typedef view_type_from_pixel<rgb12_packed_pixel_t>::type rgb12_packed_view_t; The thing is that I would now like to copy data to another rgb16_view_t... But this is not working because, in my opinion, 48 bits is between 32 and 64. I have been thinking of the solution that Christian gave me and that consist on doing something like this: T p( pData, 0 ); for( typename DST_V::y_coord_t y = 0; y < height; ++y ) { typename DST_V::x_iterator it = dst.row_begin( y ); for( typename DST_V::x_coord_t x = 0; x < width; ++x ) { color_convert( *p, *it ); ++p; ++it; } } where T is a bit_aligned_pixel_iterator and DST_V=rgb16_view_t but this can't be working on my case because channels are shifted by 4 bits ! If you could help me, I would really appreciate. I think all those cases are a good robustness test for gil. Regards, Eloi.

3 3

Trying to build docs....
by Marshall Clow 05 Jan '10

05 Jan '10

I'm trying to generate the docs for boost. I'm looking at the instructions at <http://www.boost.org/doc/libs/1_41_0/doc/html/boostbook/getting/started.html> I've installed: docbook-xml-4.2 docbook-xsl-175.2 fop-0.95 Jimi 1.0 (and done the 'move and rename' bit into fop) I believe that I've set up my user-config.jam correctly (attached). And when I try to build something, I get errors about "unable to external entity <a sourceforge URL>". If I take the <a sourceforge URL> and try to curl it, I get something that looks reasonable. Any suggestions? [ Oh, yeah: Mac OS X 10.6, gcc 4.2.1, from the trunk. ] -- Marshall > $ bjam --v2 onehtml > ...patience... > ...found 1251 targets... > ...updating 1 target... > xslt-xsltproc ../../../bin.v2/libs/fusion/doc/darwin-4.2.1/debug/fusion.html > warning: failed to load external entity "http://docbook.sourceforge.net/release/xsl/current/html/docbook.xsl" > compilation error: file /Users/marshall/Sources/boost/tools/boostbook/xsl/html-single.xsl line 15 element import > xsl:import : unable to load http://docbook.sourceforge.net/release/xsl/current/html/docbook.xsl > warning: failed to load external entity "http://docbook.sourceforge.net/release/xsl/current/html/admon.xsl" > compilation error: file /Users/marshall/Sources/boost/tools/boostbook/xsl/admon.xsl line 12 element import > xsl:import : unable to load http://docbook.sourceforge.net/release/xsl/current/html/admon.xsl > warning: failed to load external entity "http://docbook.sourceforge.net/release/xsl/current/lib/lib.xsl" > compilation error: file /Users/marshall/Sources/boost/tools/boostbook/xsl/relative-href.xsl line 12 element import > xsl:import : unable to load http://docbook.sourceforge.net/release/xsl/current/lib/lib.xsl > > XML_CATALOG_FILES="../../../bin.v2/boostbook_catalog.xml" > export XML_CATALOG_FILES > "xsltproc" --stringparam boost.libraries "../../../libraries.htm" --stringparam boost.root "../../../.." --stringparam chunk.first.sections "1" --stringparam chunk.section.depth "4" --stringparam generate.section.toc.level "4" --stringparam html.stylesheet "../../../../doc/html/boostbook.css" --stringparam toc.max.depth "3" --stringparam toc.section.depth "3" --xinclude -o "../../../bin.v2/libs/fusion/doc/darwin-4.2.1/debug/fusion.html" "/Users/marshall/Sources/boost/tools/boostbook/xsl/html-single.xsl" "../../../bin.v2/libs/fusion/doc/darwin-4.2.1/debug/fusion.docbook" > > ...failed xslt-xsltproc ../../../bin.v2/libs/fusion/doc/darwin-4.2.1/debug/fusion.html... > ...failed updating 1 target... > > mclow-mac:doc marshall$ curl http://docbook.sourceforge.net/release/xsl/current/lib/lib.xsl | bbedit > % Total % Received % Xferd Average Speed Time Time Time Current > Dload Upload Total Spent Left Speed > 100 18500 100 18500 0 0 76291 0 --:--:-- --:--:-- --:--:-- 141k >

2 1

Re: [Boost-users] [interprocess] Not implemented?
by Pascal Kesseli 05 Jan '10

05 Jan '10

Thanks for your reply. The complete system version goes as follows: Linux Moxa 2.6.23.1 #1266 Fri Nov 14 19:05:06 EST 2008 armv5teb unknown UC-8410 version 1.0 It is an embedded MOXA computer. Does this mean that shared memory objects are not fully implemented on this operating system? On some of this device's predecessors, I used standard Unix semaphores for this task, before I switched to the new boost version when the interprocess library was introduced in version 1.39.0. Would returning to this manual solution work or do semaphores not function at all on this machine? Thanks again and best regards Pascal

2 1

mpl: generate kind of permutations
by Hicham Mouline 05 Jan '10

05 Jan '10

Hello, I have 2 strictly ordered sequences of dates t1... tn, and td1...tdm, where m and n may or may not be different. We always have: t1<...<tn and td1<...<tdm. I am trying to place the (t1...tn) sequence over the (td1...tdm) sequence, and generate all these "permutations" ie . all ti < td1 or. all ti < tn = td1 or. all ti except tn < td1 and (tn >td1 or tn=td2 or td3>tn>td2 or tn=td3 or ....tn>tdm-1 or tn=tdm or tn>tdm) or. t1...tn-2 < td1 and tn-1=td1 and (td1<tn<td2 or tn=td2 or tn>td2 or tn=td3 or ....tn>tdm-1 or tn=tdm or tn>tdm) or. t1...tn-2 < td1 and td1<tn-1<td2 and (the same) or. t1...tn-2 < td1 and td2=tn-1 and (the same) ...etc...etc or. t1 < td1 and distribute the rest of t_i_s or. t1 = td1 and distribute the rest of t_i_s or. td2> t1 > td1 and distribute the rest of t_i_s etc or. tdm< t1< t2< ...<tn I hope the description is clear enough even though it is not formal at all. The problem is like distributing n balls over m-1 buckets or to the left or to the right of all buckets or exactly at bucket borders (with never more than 1 t_i on any bucket order) In the runtime world, the pseudo-code would look like: calling i0: number of t_is in the 0th bucket (the one to the left) calling i1: number of t_is in the 1th bucket (between td1 and td2) ... calling im-1: number of t_is in the (m-1)th bucket (between tdm-1 and tdm) calling im: number of t_is in the mth bucket (>tdm) j1 : number of t_is = td1 (0 or 1) ... jm: number of t_is = tdm (0 or 1) for (i0 = n; i0>=0; --i0) for (i1 = n-i0; i1>=0; --i1) for (i2 = n-i0-i1; i2>=0; --i2) ... for (im = n-i0-i1-im_1; im>=0; --im) ... 0. The pseudo code is incorrect here. I need to account for the j_s as well. 1. I would normally think of using recursion to generate the serie of inner-loops. 2. I believe there must be a purely iterative solution. 3. This problem seems familiar in that perhaps one of the STL algorithm does something similar 4. n and m are never bigger than 4 or 5. Though already, that yields a large number of permutations. I haven't figured out yet the exact number. So I think compilation will be slow but not unbearably (it's a guess). I am interested in doing this in compile-time though. I encode dates as an ints. I would start with the td dates the mpl::vector_c< 20100109, 20100125, 20100322, 201004228 > which gives m=4. Then I would set n to 3 for e.g. Then I would ask the mpl metafunction to return all the possible combinations. That would be a vector or vectors? vector_c< permut1, permumt2, ... permut_last > For each vector of the permut_s, I would have a runtime function to run (chooses dates and distribute them as provided by the vector). Regards,

3 3

[interprocess] Not implemented?
by Pascal Kesseli 05 Jan '10

05 Jan '10

Hi folks Using boost 1.39.0, I receive the following error message when running my program on the target system: terminate called after throwing an instance of 'boost::interprocess::interprocess_exception' what(): Function not implemented Aborted My program uses boost named mutexes. How do I have to interpret this exception? The target is a MOXA embedded UC-8410 computer, featuring a GCC 4.2 platform. Thanky for any suggestions and best regards Pascal _________________________________________________________________ Samichlaus du liebe Maa, hesch dis Hotmail hüt scho gha? Gratis Geschenk runterladen! http://www.microsoft.com/switzerland/windows/de/windowslive/hotmail_bl1/hot…

2 1

Re: [Boost-users] [unit_test_framework] checkpoints for BOOST_CHECK*?
by rozelak＠volny.cz 04 Jan '10

04 Jan '10

> > rozelak(a)volny.cz wrote: > >> rozelak(a)volny.cz wrote: > >> > >>> Is there a way how to achieve what I am looking > >>> for? > >>> >>> > >> Use BOOST_TEST_MESSAGE and set log level to message > >> >> > > > > Yes, I have read about this option and it works - > > thank you. I consider > > > it as not perfect, but working solution. > > > > However, I want to ask, if the idea of checkpoints > > could be generalised > > > and used during check failure printing as well (as > > has been suggested > > > in the example in my first post). > > I don't know boost source in depth, but it seems > > that BOOST_TEST_CHECKPOINT > > > just calls boost::unit_test::unit_test_log.set_checkpoint(...), > > > > > > > > which > > > stores the checkpoint info in unit_test_log (which > > is supposed to be > > > singleton). > > Would, therefore, be possible to print the checkpoint > > stored also when > > > test failure is printed? May I be helpfull somehow > > to do this? > > > > > Or is it against the idea of the checkpoints? > > > > I think the issue is that BOOST_CHECK and friends set > a checkpoint. > I see, you are right. It is set by BOOST_TEST_PASSPOINT() macro. Well, do you think, that it would be beneficial to define new macro, e.g. BOOST_TEST_FAILRECORD (please suggest a better name, if you have one)? The macro would be implemented in the way very similar to BOOST_TEST_CHECKPOINT, but the data "recorded" by the macro would be printed only when a testcase fails or an unexpected exception occurs. It also has the advantage, that when not used, it does not affect the current behaviour of boost. Is there a change to accept it in the boost test framework, if I try to implement it? Thank you for suggestions/comments, Dan

3 7

Using atan2 in boost::units
by Brandon Olivares 04 Jan '10

04 Jan '10

Hi, I need to use the boost::units::atan2() function in boost/units/cmath.hpp, but am having trouble figuring out how. Do the units have to be dimensionless, or can they be any unit? I have two variables, x_ and y_ which are of a scaled unit type, called light_second_unit. I'm calling it like this: boost::units::atan2(y_, x_); I get this error: ./physics/cartesian_coordinate_system/point.cpp:85: error: no matching function for call to `atan2(boost::units::quantity<boost::units::astronomical::light_second_unit, double>, boost::units::quantity<boost::units::astronomical::light_second_unit, double>)' If it has to be dimensionless, how do I convert them to a dimensionless quantity? Thanks, Brandon -- www.creatorspiritus.com Blog and forum discussing living a Christ-centered life Follow me on Twitter: http://twitter.com/devbanana Follow creatorspiritus.com on Twitter: http://twitter.com/creatorspiritus

3 6

[Wave] Modifying the way function-like macros are interpreted
by fractured＠modularsystems.sl 04 Jan '10

04 Jan '10

I am trying to use boost::wave as a preprocessor for LSL, a language used in the virtual world Second Life. There is a data type there for collections of data, but it can only be accessed through functions such as llList2String(list,iterator); I wanted to define macros that would allow me to do something like #define x[y] llList2String(x,y), however, wave as it stands does not seem to have any immediately obvious way to do this.

4 6

[test] 64bit test case deletion errors again
by Alastair Rankine 04 Jan '10

04 Jan '10

This issue has been discussed before on this list, see the thread at gmane.comp.lib.boost.user/51628 I can confirm that the fix in 1.41 does indeed work for most regular test cases. However I believe that there is still an issue when using fixtures. Below is a sample app which crashes for me when compiled with gcc on MacOS X 10.6.2. If someone could confirm that I'm not doing anything particularly stupid, that would be appreciated. #define BOOST_TEST_MODULE autoreg test #include <boost/test/unit_test.hpp> struct Fixture { Fixture() : n(1) {} unsigned n; }; BOOST_AUTO_TEST_SUITE(AutoregTest) BOOST_FIXTURE_TEST_CASE(Test1, Fixture) { BOOST_CHECK_EQUAL(n, 1); } BOOST_FIXTURE_TEST_CASE(Test2, Fixture) { BOOST_CHECK_EQUAL(n, 1); } BOOST_AUTO_TEST_SUITE_END() % ./autoregtest Running 2 test cases... *** No errors detected autoregtest(52796) malloc: *** error for object 0x30001002006d0: pointer being freed was not allocated *** set a breakpoint in malloc_error_break to debug Breakpoint 1, 0x00007fff818f5951 in malloc_error_break () (gdb) bt #0 0x00007fff818f5951 in malloc_error_break () #1 0x00007fff8181f143 in free () #2 0x000000010006dc62 in ~framework_impl [inlined] () at /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_ports_devel_boost/work/boost_1_41_0/boost/test/impl/framework.ipp:133 #3 0x000000010006dc62 in __tcf_1 () at framework.ipp:225 #4 0x00007fff8182b334 in __cxa_finalize () #5 0x00007fff8182b24c in exit () #6 0x00000001000031e7 in start ()

1 0

asio - interrupting a long process running in server
by Lloyd 04 Jan '10

04 Jan '10

Hi, I am using ASIO to create a server. The clients sends command to the server, the server executes it in it and sends back the result to the client (here the client remotely controls the server). Lets assume that one command implementation at the server side contains only an infinite loop, as this wastes the server resource, the client need to "interrupt" or "terminate" the execution of the infinite loop in the server. Is there any mechanism available in ASIO to accomplish this? What I am doing now is, this long job is created as a thread, collect the thread's id and store it in a container. Then this id is send to the client, if the client wants to interrupt the thread, it sends another command (STOP_THREAD ) with this thread id, the server looks for the thread running with this id, and interrupts it. At the end of execution thread, the thread id has to be removed from the container (thread id,pointer to thread). But this makes it very complicated, and it is not working right. So I am looking for a solution available with ASIO... Please share your ideas... Thanks, Lloyd ___________________ Happy New Year 2010

1 0

[BGL] Free graph - double free or corruption
by Matthieu BOUSSARD 04 Jan '10

04 Jan '10

Hello, I'm using BGL to construct an exploration graph. At each decision step I have to create a new graph. It seems that I'm not freeing correctly the memory, causing a memory leak. The name of a node in the graph is a pointer to a structure, typedef adjacency_list <listS,vecS, directedS, property<vertex_name_t, state_t *, property<vertex_prop_t,s_prop> >, property<edge_weight_t, float, property<edge_name_t,int> > >Graph; I also use a map to find the vertex according to a given state typedef std::tr1::unordered_map<state_t *, vertex_t> state_vertex_map_t; In the destructor of my planner object, every time I try to free anything (from the map or from the vertices of the graph) I'v got an "double free or corruption" error. Could you help me to find the way to free all the memory ? Thank you. Best regards Matthieu

2 4

Re: [Boost-users] Defining Statements with Proto
by David A. Greene 04 Jan '10

04 Jan '10

On Sunday 03 January 2010 18:55:36 you wrote: > On Sunday 03 January 2010 12:15:59 you wrote: > > > There were a few other small problems, but on the whole, you were > > > pretty close. > > You moved the code around a bit so the diff isn't trivial. What were the > other small problems? Ah, I found one. Duplicated template parameters for the expression rules. -Dave

1 0

[test] Unexpected Exceptions
by Florian Winter 04 Jan '10

04 Jan '10

Hi, What is the intended behavior of Boost.Test if an unexpected exception is thrown? Example: void do_not_throw() throw() { throw "unexpected"; } Some of my applications using Boost.Test seem exit with an error code of 0 (successful) if an unexpected exception is thrown. Is this the intended behavior? Best regards, fw

3 5

[UTF] Accessing global fixture members within tests
by Simon Pickles 04 Jan '10

04 Jan '10

Hi I am using the Unit Test Framework to test my use of boost.python. I am using a global fixture to initialize the python intepreter a single time (rather than before each test). My problem is I would like to use the Python initialising class in my tests, and I can't figure out how to do this. I don't want to use normal fixtures since each test would call initialize the interpreter, causing problems. Here is what I want to do: ////////////////////////////////////////////////////// #include <boost//python.hpp> #include <boost//test//unit_test.hpp> #include "..//pythonManager//pythonManager.h" using namespace boost; // First make a global fixture which sets up python interpreter struct PythonManagerFixture { PythonManagerFixture() { pm.Init(); } PythonManager pm; }; // Set up suite BOOST_AUTO_TEST_SUITE(TestPythonManager) // Only init python interpreter once BOOST_GLOBAL_FIXTURE(PythonManagerFixture); BOOST_AUTO_TEST_CASE(TestInit) { // Get python globals dict python::dict globalDict = python::extract<python::dict>(pm.m_pyGlobals); // COMPILE ERROR 'pm' : undeclared identifier // Has logging module been imported? BOOST_REQUIRE(globalDict.has_key("logging")); } BOOST_FIXTURE_TEST_CASE(TestLogCallback, PythonManagerFixture) // RUNTIME ERROR pm.Init called a second time - bad { function<void(const string&, const string&)> f = pm.GetLogFunction(); // pm is in scope though.... BOOST_REQUIRE(f); } // End the suite BOOST_AUTO_TEST_SUITE_END() /////////////////////////////////////////////////////////////////// I need a way to supply the PythonManager instance, pm, to the test cases. Thanks for any advice Simon

2 3

alignment question
by Slawomir Lisznianski 03 Jan '10

03 Jan '10

I'm having trouble implementing properly aligned storage. The following example shows conceptually what I'm trying to do, but it's NOT a correct way of doing it. template <typename T> struct envelope { size_t len = sizeof(T); // <--- compiler may pad here, hence this doesn't work... T body; }; So, given an arbitrary type T, we want to create an envelope that will have the size of T immediately before it, with no padding between len and body. The user would then be able to: envelope<MyStruct> env; env.body.x = ... send(env); I'm curious if anyone has run into something similar and have a clean solution. Our code currently has to defer to temporary buffers and memcpy'ing data back to it. Thanks, Slaw

3 2

Re: [Boost-users] regex
by ninti＠internode.on.net 03 Jan '10

03 Jan '10

Thanks for your help, linking was exactly the problem. As you can no doubt tell, I'm new not only to Boost but C++ in general. If you have the patience for one more newbie question ... I have the following code (in a FastCGI++ web application). The intention is to sanitise user input by allowing only alphanumeric characters, spaces, full stops and hyphens, simply replacing everything else with underscores: std::string input_name = environment.posts["name"].value; std::string filler = "_"; boost::regex re("[^a-zA-Z0-9\s.-]"); std::string output_name = boost::regex_replace(input_name, re, filler); Compilation fails with the following error: contact.cpp: In member function ‘virtual bool Contact::response()’: contact.cpp:48: error: unknown escape sequence '\s' The regex works fine with the \s removed (but of course replaces all spaces with underscores). These alternatives compile OK but don't work (program still replaces all spaces with underscores): [^a-zA-Z0-9\\s.-] [^a-zA-Z0-9[:space:].-] I know this is going to be something pretty basic, but how can I include whitespace - non-breaking, breaking, tabs, newlines, the lot - in the regex above as symbols to NOT be replaced? Thanks again Mike

2 1

regex
by ninti＠internode.on.net 03 Jan '10

03 Jan '10

Boost 1.40 on CentOS 5.3 64bit: I have #included <boost/regex.hpp> in a file but get the following compile time error: /tmp/ccXM0HkR.o: In function `boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::assign(char const*, char const*, unsigned int)': contact.cpp:(.text._ZN5boost11basic_regexIcNS_12regex_traitsIcNS_16cpp_regex_traitsIcEEEEE6assignEPKcS7_j[boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::assign(char const*, char const*, unsigned int)]+0x27): undefined reference to `boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::do_assign(char const*, char const*, unsigned int)' collect2: ld returned 1 exit status This file definitely exists at /usr/include/boost/regex.hpp. Another Boost #include in the same file (<boost/date_time/posix_time/posix_time.hpp>) produces no error. 'ldconfig -p' lists libboost_regex.so. What am I doing wrong? Thanks Mike

2 1

regex again
by ninti＠internode.on.net 03 Jan '10

03 Jan '10

Sorry if this is double-posted, but my webmail has stripped angle brackets and their contents out of my previous email ... Boost 1.40 on CentOS 5.3 64bit: I have #included boost/regex.hpp in a file but get the following compile time error (mangled by webmail): /tmp/ccXM0HkR.o: In function `boost::basic_regex boost::cpp_regex_traits > >::assign(char const*, char const*, unsigned int)': contact.cpp:(.text._ZN5boost11basic_regexIcNS_12regex_traitsIcNS_16cpp_regex_traitsIcEEEEE6assignEPKcS7_j[boost::basic_regex boost::regex_traits > >::assign(char const*, char const*, unsigned int)]+0x27): undefined reference to `boost::basic_regex > >::do_assign(char const*, char const*, unsigned int)' collect2: ld returned 1 exit status The important bit appears to be 'undefined reference to boost__basic_regex' etc This file definitely exists at /usr/include/boost/regex.hpp. Another Boost #include in the same file (boost/date_time/posix_time/posix_time.hpp) produces no error. 'ldconfig -p' lists libboost_regex.so. What am I doing wrong? Thanks Mike

1 0

[Accumulators] Evaluation of Expression Tree
by Horst S 02 Jan '10

02 Jan '10

Hello! Given an expression tree over mathematical expressions like 'sum', 'min', 'product', etc. and a fixed number of variables X1..Xn like the following, would it be possible to use BOOST Accumulators to re-evaluate the root evaluation function based on changes in the values of the variables efficiently? sum / \ / \ min product / \ / / \ X1 X2 X3 X4 Note that variables appear only as leaf nodes. Obviously, one can flatten the tree to the following expression: RootValue = sum( min(X1,X2), product(X2,X3,X4) ) Assuming that the current values of the variables are X1=3, X2=4, X3=2, X4=2 then RootValue evaluates to 19. Now setting X2=2 results in 10. The main objective is to minimize the cost of computing the value of the root evaluation function when changing the value of a subset of variables in X1..Xn. One can assume that the subset of variables that changes each time is rather small (typically only one) and the size of the expression tree is rather large. Are BOOST Accumulators a suitable way to model such kind of expressions and the corresponding incremental evaluation? Thanks!

2 1

[boost] [program options] Recommended way to specify a list of strings
by Sanjit Jhala 01 Jan '10

01 Jan '10

Say I want to specify a list of servers via the boost program_options library. Something like: --total-servers N --server srv1 --server srv2 ... --server srvN How would I do this? It doesn't look like multitoken would work for this case, so would it be better for me to just have something like: --total-servers N --serverlist srv1 srv2 ... srvN and use a single string to store the serverlist and parse it myself? -Sanjit

4 6