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

Jam: Building libraries that use the static runtime library
by Dale Peakall 13 Mar '02

13 Mar '02

I need to build libraries that link against the static runtime library: I tried the following Jam invocation: jam -sBOOST_ROOT=. -sTOOLS="msvc-stlport" -sBUILD="debug release <runtime-link>static" But jam balks and reports: Error: <libs!python!build>libboost_python_static: target requirements conflict for requested build { <runtime-link>dynamic <target-type>LIB <runtime-link>static } Help! So, I've obviously got it wrong (btw I'm not interested in building python library just threads and regex). TIA. - Dale.

2 1

AIX and Boost 1.27.0 (Does not compile)
by samsrd 12 Mar '02

12 Mar '02

Has anyone tried to compile Boost 1.27.0 (or any version for that matter) on IBM AIX using the xlC compiler (IBM Visual Age). There are many compile errors (usually template related) along the lines of: "../../../../ThirdParty/boost/regex/detail/regex_cstring.hpp", line 77.40: 1540- 0152 (W) A template dependent name that is a type must be qualified with "typena me". "../../../../ThirdParty/boost/regex.hpp", line 528.55: 1540-0152 (W) A template dependent name that is a type must be qualified with "typename". "../../../../ThirdParty/boost/regex.hpp", line 532.13: 1540-1628 (S) Every templ ate parameter for a constructor template must be used in the parameter list of t he constructor. "../../../../ThirdParty/boost/regex.hpp", line 532.34: 1540-0152 (W) A template dependent name that is a type must be qualified with "typename". "../../../../ThirdParty/boost/regex.hpp", line 744.13: 1540-0152 (W) A template dependent name that is a type must be qualified with "typename". "../../../../ThirdParty/boost/regex.hpp", line 938.4: 1540-0152 (W) A template d ependent name that is a type must be qualified with "typename". "../../../../ThirdParty/boost/regex.hpp", line 438.27: 1540-0130 (S) "def_alloc_ param_traits" is not declared. "../../../../ThirdParty/boost/regex.hpp", line 1414.1: 1540-0062 (S) The incompl ete class "match_results" must not be used as a qualifier. "../../../../ThirdParty/boost/regex.hpp", line 1455.35: 1540-0140 (S) The text " &" is unexpected. "match_results" may be undeclared, ambiguous or may require " typename" qualification. "../../../../ThirdParty/boost/regex.hpp", line 444.7: 1540-1303 (I) "template cl ass match_results" is not visible. @

2 1

thread::yield() vs Sleep()ing
by simonwlb 10 Mar '02

10 Mar '02

I'm using thread::yield() in my multi-threaded cross-platform application to avoid "CPU burn". This is because I have threads that are in an infinte loop in order to do some polling. There are no mutexes to wait on in the loop. void X::threadFunction() { for(;;) { if( m_x->isReady() ) doSomething(); . . . boost::thread::yield(); } } If the answer is "you should almost never poll like that - try to use mutexes etc. as much as possible" please say so (I'm a newbie to threads)! In any case, if I choose to have a thread in an infinite loop anyway, I have three options (assuming I have no mutexes available to me): 1. Don't yield or Sleep at all. 2. Sleep for a while - say for example, 10 ms. 3. Yield. Option 1 is clearly wrong. The CPU usage will go to 100% and the whole machine may slow down to an unusable speed. Option 2 works well in my emperical tests. The thread loop executes plenty fast enough and my Windows CPU meter shows virtually 0% CPU usage. 10ms is a bit of an arbitrary figure though - My code needs to work on a variety of software and hardware platforms, so the ideal number is hard or impossible to know. Option 3 also works well in my emperical tests and seems like the neat solution (no arbitrary waiting period). My application and the system as a whole both run smoothly and fast. The only irksome thing is that the CPU meter shows 100% usage. That's a bit alarming for me and the users of my software. My questions: 1. Should thread loops without logical wait conditions be avoided as much as possible? Is that the problem? 2. Should I just yield() or should I pause for a time period? 3. How do you select a time period to wait? Simon Bailey

4 4

Boost Building
by karl_m_au 09 Mar '02

09 Mar '02

Hiya I'm trying to build boost 1.27.0 on a Debian linux system. I've built the boost version of jam. Both g++ 2.95.3 and g++ 3.0.3 are installed. I'm pretty sure I've made the appropriate changes to get g++ 3.0 to be used in the build process. I've used various versions of the command line recommended in the documentation. Very few of the builds work, becuse of this: g++-3.0 -c -Wall -ftemplate-depth-100 -I"libs/regex/build" -o "libs/regex/build/bin/libboost_regex.a/gcc-3/debug/c_regex_traits.o" "libs/regex/build/../src/c_regex_traits.cpp" Somehow the <include> directive isn't working. Any clues? TIA Karl M

3 4

Boost python
by andrewperella 08 Mar '02

08 Mar '02

I am having great trouble with boost python whenever I do anything slightly more compilcated than the examples. I have a wrapped a Vector3 class with 3 floats x,y,z. In python: a=m2.Vector3(1,2,3) a.x 1.0 a.x=100 a.x 100.0 This is correct - ie I can disply and modify a.x Unfortunatedly if wrap another class - Camera which contains a member variable : Vector3 pos then I cannot set camera.pos.x Looking in the debugger, I seem to be getting a copy (in fact sometime 3 or more copy constructors are called) of pos and I am modifying this temporary copy! Help. Surely I can wrap no pointer member variables? Regards, Andrew

1 0

1.27 BGL and VC++7.0 compilation problem
by elgatomoses 07 Mar '02

07 Mar '02

I am having trouble compiling the Graph library with VC++7.0 (.NET) - I'm getting "INTERNAL_COMPILER_ERROR". The problem has to do with the is_array definition in composite_traits.hpp. Has anyone else encountered this problem - perhaps there are some compiler switches that I can set... TIA Mark Kruger

2 1

Boost.build question: Adding a new unit-test rule
by Helge Penne 04 Mar '02

04 Mar '02

Hi. I'm working on a project where some of the unit test are very inconvenient to implement with CppUnit and Mock classes or stubs alone, but would be very simple to implement by having them print messages to standard output, pipe this to a log file and diff this against a reference file. I added the following to our Jamrules file (most of it is stolen from the standard unit-test rule and run action): rule unit-test-diff { local files = [ exe $(<) : $(2) : $(3) : $(4) : EXE ] ; type-DEPENDS test : $(files) ; local file ; for file in $(files) { run-and-diff $(file) ; } } actions run-and-diff { $(SHELL_SET)PATH=$(RUN_PATH) $(SHELL_EXPORT)PATH $(<) > $(<:D=:S=.log) diff $(<:S=.log) $(<:S=.ref) # diff $(<:D=:S=.log) $(<:D=:S=.ref) } This runs the unit test program, pipes the output to <targetname>.log and diffs agains <targetname>.ref. There is only one problem: I cannot get Boost.build to locate these files in the correct folder. The Jamfile for the unit test is located in the same folder as the unit test source code. This is the only natural place for the ref-file as well. The action above has three different alternatives for the actual diff operation. The first uses the full path of the executable, but replaces the last part of the file name with .log and .ref. This places them in the binary folder of the unit test program, which isn't not very convenient. The second strips away the path, but since the currect folder always is the project root (not the folder where the Jamfile is), this places the log and ref files on the project root, which is almost as bad as the first alternative. I also experimented briefly with :G but this has no effect as the file path did not seem to contain grist. One possible solution might be to add a path feature to the target and set this to the reference file name, but I'm not quite sure how to implement this for rules that are not part of a tools.jam-file I really need the unit test rule to place these files in the folder where the rest of the test source code is. Does anyone know how to do this? - Helge Penne

3 2

(no subject)
by Hook.RobertA(CC) 04 Mar '02

04 Mar '02

I put the following up at source forge support, then realised that doesn't seem to be likely to get a response.... I would have put this through the yahoo group, but I'm stuck behind a mind-bogglingly repressive firewall. SO I'll scream for help here, bearing in mind the dictum that in cyberspace no-one can hear you scream. I have download 1.27.0, and am attempting to build and use Boost under HP-UX 11.00. There are two problems, and I cannot tell if they are related or not. Firstly, the install instructions along the lines of jam -sBOOST_ROOT=. -=sTOOLS="hpux" simply produce the following (after I bodgied up a hpux-tools.jam) +: line 1304: yacc stack overflow at keyword } don't know how to make all ...found 1 target... ...can't find 1 target... after smacking my head on the desk for several days, I've given up on that one and attempted to build boost_1_27_0/libs/tuple/test/io_test.cpp using aCC -c +z +p $APP_INCLUDE -o io_test.o io_test.cpp (where APP_INCLUDE is -I <yada/yada/yada> as appropriate) which results in "tuple_basic.hpp", line 290 # Cannot generate a specialization of incomplete template struct STATIC_ASSERTION_FAILURE<(bool)0>. BOOST_STATIC_ASSERT(length<cons>::value == 2); // check length = and errors cascading after that. Looking at various messages and comments, I believe that someone somewhere has managed to get boost working under HP-UX, but I'm dammned if I can. Help! Oh yes, by the way, I'm stuck behind a repressive firewall and cannot even *read* the message archive at Yahoo, let alone subscribe / post! (Puts head down on desk and begins sobbing quietly) -- Robert Hook - DCC Analyst/Programmer Polaris Group, Information Systems Branch Queensland Police Service Hook.RobertA(a)police.qld.gov.au PH: 07 3364 3807 ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify securityscanner(a)police.qld.gov.au. This footnote also confirms that this email message has been checked for the presence of computer viruses. **********************************************************************

1 0

[Fwd: compile time container type identification]
by hicks 01 Mar '02

01 Mar '02

Hello I need to identify if a container is of a certain type. Specifically, from inside the code for container template <class T> Cont, I want to identity if T is itself of type Cont<U> for some U. (The type of U does not matter.) I couldn't see anything in *_traits.hpp that looks applicable. Any advice or references would be appreciated. Craig Hicks KGK Tokyo [Non-text portions of this message have been removed]

1 0

RE: [Boost-Users] Linking an Archive
by Tom Matelich 27 Feb '02

27 Feb '02

The first getting_started1.so contains all the symbols from all of the objects linked into it. The second one only contains the symbols from getting_started.o and the symbols from libboost_python.a which satisfy unsatified symbols of getting_started1.o. Linking shared library is like linking an executable, it only takes the symbols from archives that it needs to satisfy itself. The first question is why you want libboost_python.a. If it is a valid reason, then you need to find a way to get all the code you need out of it into the shared lib. The first would be to use ar to get all the objects out, then take those and put them into the shared library. The second would be to keep adding unsatified symbols to getting_started.o until getting_started1.so has everything it needs. Tom > -----Original Message----- > From: doak87508 [mailto:jdoak@lanl.gov] > Sent: Wednesday, February 27, 2002 10:21 AM > To: Boost-Users(a)yahoogroups.com > Subject: [Boost-Users] Linking an Archive > > > I am playing around with using an archive to created a shared library > (.so file) instead of linking all the individual objects. I have > success when I link individual objects: > > % g++ -shared classes.o conversions.o errors.o extension_class.o > functions.o init_function.o module_builder.o objects.o types.o > cross_module.o getting_started1.o -o getting_started1.so > > % python > 1;31mh[1] >>> import getting_started1 > 1;31mh[1] >>> getting_started1.greet() > 'hello, world' > > Now, I create an archive: > > % ar r libboost_python.a classes.o conversions.o errors.o > extension_class.o functions.o init_function.o module_builder.o > objects.o types.o cross_module.o > > Then, I link that archive into the shared library: > > % g++ -shared libboost_python.a getting_started1.o -o > getting_started1.so > > % python > 1;31mh[1] >>> import getting_started1.so > Traceback (most recent call last): > File "<stdin>", line 1, in ? > ImportError: ./getting_started1.so: undefined symbol: > add__Q35boost6python19module_builder_basePQ45boost > 6python6detail8functionPCc > > I have also tried linking as follows with no change in result: > > % g++ -shared -L. -lboost_python getting_started1.o -o > getting_started1.so > > Does someone know how to get linking working with an archive file? > > JD > > > > Info: <http://www.boost.org> > Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> > Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > > Your use of Yahoo! Groups is subject to > http://docs.yahoo.com/info/terms/ > >

2 1