- Boost - lists.boost.org

3 important Intel regressions and some other stuff
by Stefan Slapeta 03 Aug '04

03 Aug '04

The following regressions seem to be special to Intel 8: - numeric/uBLAS The only way to solve this seems at the moment not to use the "Multithreaded Debug DLL" runtime. When we can isolate the problem, Intel should investigate. - Python The slice test case is failing! Any ideas? - random error: no operator "<<" matches these operands I've not heard any comment in the last days about that. Any ideas? Aleksey, could we have a link to http://boost.sourceforge.net/regression-logs/developer/python.html on the boost regression site? I can't find any link to this page anywhere. I'm now going to leave now and I'll return on Sunday. I don't know if I can manage to get the regressions being run automatically, otherwise you get the next results for Intel 8 + CW 9 on Sunday. Stefan

5 16

[multi_index] spurious error in CW 9
by Joaquín Mª López Muñoz 03 Aug '04

03 Aug '04

This test occasionally fails: http://tinyurl.com/72xcb but there's no error output or any failure indication, nothing. The problem shows only at times, so I guess it'll have to do with the compiler being low on resources, but couldn't say for sure. Any idea? Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

4 3

[regression] Discrepancy with display and results...
by Rene Rivera 03 Aug '04

03 Aug '04

I just noticed a discrepancy between the regression tables, the MetaComm one's, and the output reported for the failure. If one looks at the Spirit summary: http://www.boost.org/regression-logs/cs-win32_metacomm/developer/spirit.html Boost regression: spirit/CVS main trunk There are two failures, action_tests and action_tests_debug. But if one looks at either for the failure, you see that there is no failure: Boost regression - test run output: spirit - action_tests / cwpro8 Compiler output: call "C:\Program Files\Metrowerks\CodeWarrior\Other Metrowerks Tools\Command Line Tools\cwenv.bat" -quiet mwcc -maxerrors 5 -maxwarnings 20 -c -warn on,nounusedexpr,nounused -cwd include -DNOMINMAX -nowraplines -lang c++ -g -O0 -inline off -prefix UseDLLPrefix.h -runtime dmd -iso_templates on -I"C:\Users\Administrator\boost\main\results\bin\boost\libs\spirit\test" -I"..\libs\spirit" -I".." -I- -I"C:\Users\Administrator\boost\main\boost" -o "C:\Users\Administrator\boost\main\results\bin\boost\libs\spirit\test\action_tests.test\cwpro8\debug\unit_test.obj" "..\libs\spirit\test\./actor/unit_test.cpp" Linker output: call "C:\Program Files\Metrowerks\CodeWarrior\Other Metrowerks Tools\Command Line Tools\cwenv.bat" -quiet mwld -search -maxerrors 5 -maxwarnings 20 -export dllexport -nowraplines -g -subsystem console -o "C:\Users\Administrator\boost\main\results\bin\boost\libs\spirit\test\action_tests.test\cwpro8\debug\action_tests.exe" @"C:\Users\Administrator\boost\main\results\bin\boost\libs\spirit\test\action_tests.test\cwpro8\debug\action_tests.CMD" Run output: Running 14 test cases... *** No errors detected EXIT STATUS: 0 -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com - 102708583/icq

3 7

Another project based on Boost Graph
by Jonah Beckford 03 Aug '04

03 Aug '04

I would like to thank the Boost Graph implementers (and the general Boost community) for coming up with a great set of patterns for dealing with graphs. I have started something similar for Scheme (a Lisp dialect), and would like to acknowledge where the great ideas came from. Keep up the good work! Jonah http://www.nongnu.org/rgraph/ README excerpt: Rooster Graph is a graph library for Scheme. Patterned after the popular Boost Graph Library for C++, it allows for a very clean separation between the graph container and graph algorithms. It does not have the complete functionality of the Boost Graph Library, but it does go a bit further by having both strict- and lazy-evaluated methods; the user of the graph has pause/continue/stop control and is not constrained by memory. Rooster Graph uses Scheme macros to get the benefit of speed while having generic operations. It has only been implemented on CHICKEN Scheme, although porting it to other Scheme implementations would be quite easy.

1 0

[program_options] Current status
by Vladimir Prus 03 Aug '04

03 Aug '04

I've leaving on vacation until Aug 23, so I though I'd explain the current release status of program_options. I believe that most failures are addressed. The Boost-wide table shows a lot of yellow cells, but I'm sure I've added explicit markup for most of them, and indeed the results for win32 are much better. The only two unexpected yellow cells are for intel- win32- 7.1- vc6- stlport- 4.5.3 at http://boost.sourceforge.net/regression-logs/cs-win32_metacomm/developer/pr… These are fresh problem and I'm not sure how I've missed them, or when they are introduced. In any case, it's too late to fix them. Another problem is that I still not sure about Darwin. All tests fail because of Boost.Test problem, and I'm not sure if Gennadiy is going to do anything about it. I think this is very bad situation, and would say that reverting the problematic change in Boost.Test is an option, but can't press for it. Apart from the above, during last week I've applied numerous editorial changes to the docs (big thanks to Charles Brockman), revived the remaining examples, and added new sections to the "howto" section, so I hope docs are in a good shape as well. If during the release process anybody would want to fix something, feel free to commit the change, making sure it does not breaks anything else, and dropping me a note by personal email. Thanks again to everyone who already sent patches for the library! - Volodya

2 3

RE: problem with subexpressions
by Jadhav, Rahul (Rahul) 03 Aug '04

03 Aug '04

Thanks a lot for your reply. Problem solved. Rahul Jadhav.

1 0

Numeric Interval failing with __USE_ISOC99 environment (plus fix)
by Michael Stevens 03 Aug '04

03 Aug '04

I am just testing Boost uBLAS and Interval on an X86_64 architecture machine. The compilation environment (GCC on Linux) is defining __USE_ISOC99, and not __i386__ which seems reasonable! Looking at 'hw_rounding.hpp'' I have confirmed that on this environment the following code branch is being taken. # include <boost/numeric/interval/detail/c99_rounding_control.hpp> With this setup 'test_float.cpp' is failing badly when compiled with the following errors: In file included from /boost_CVS/boost /numeric/interval.hpp:18, from /boost_CVS/libs/ numeric/interval/test/test_float.cpp:11: /boost_CVS/boost/numeric/interval/rounding.hpp: In instantiation of `boost::numeric::interval_lib::save_state<boost::numeric::interval_lib::rounded_arith_opp<float, boost::numeric::interval_lib::rounding_control<float> > >': /boost_CVS/boost/numeric/interval/hw_rounding.hpp:40: instantiated from here /boost_CVS/boost/numeric/interval/rounding.hpp:78: error: no type named `rounding_mode' in `struct boost::numeric::interval_lib::rounded_arith_opp<float, boost::numeric::interval_lib::rounding_control<float> >' /boost_CVS/boost/numeric/interval/rounding.hpp: In instantiation of `boost::numeric::interval_lib::save_state<boost::numeric::interval_lib::rounded_arith_opp<double, boost::numeric::interval_lib::rounding_control<double> > >': /boost_CVS/boost/numeric/interval/hw_rounding.hpp:45: instantiated from here etc.. Could it be that 'struct c99_rounding_control' should be inheriting from 'c99_rounding'. This seems to fix the problem although I am not sure it is the correct solution! Michael ___________________________________ Michael Stevens Systems Engineering Navigation Systems, Estimation and Bayesian Filtering http://bayesclasses.sf.net ___________________________________

1 0

regressions in como-win32-4.3.3-vc7.1
by JOAQUIN LOPEZ MU?Z 03 Aug '04

03 Aug '04

Please see http://tinyurl.com/5f7ba "C:\Comeau\libcomo\cnames\cstring", line 31: error #247: function "std::strlen" has already been defined extern "C++" inline size_t strlen(const char *arg) { return ::strlen(arg); } ^ "C:\Comeau\libcomo\cnames\cstring", line 31: error #247: function "std::strlen" has already been defined extern "C++" inline size_t strlen(const char *arg) { return ::strlen(arg); } ^ 2 errors detected in the compilation of "..\libs\multi_index\test\test_basic_main.cpp". The problem has nothing to do with Boost.MultiIndex, actually, as test_basic_main.cpp does not do anything more than including <boost/test/included/test_exec_monitor.hpp> These tests used to pass one or two days ago. I've inspected Boost.Test for recent changes and found none. Other libraries are affected by this problem: algorithm/minmax algorithm/string program_options ... Thanks Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

[preprocessor] Digital Mars port
by Daniel James 03 Aug '04

03 Aug '04

I've ported the preprocessor library to the Digital Mars preprocessor. The patch is at: http://myweb.tiscali.co.uk/calamity/patches/preprocessor-1.patch.gz With it, the only test which doesn't pass is the BOOST_PP_LINE test in debug.cpp, because the preprocessor has problems with macro expansion in #line statements. Would it be possible to add this to CVS after the release? Unless I've made a mistake, it won't interfere with any other preprocessors. The patch works around two bugs in the preprocessor. First, in a statement like this: #define FOO(pred, x) BAR(pred(x)) pred(x) won't get expanded. There are several ways to work around this, the one I've mostly used is to write: #define FOO(pred, x) BAR(pred##(x)) You can also use: #define CALL(pred, x) pred(x) #define FOO(pred, x) BAR(CALL(pred, x)) And, with the patch: #define FOO(pred, x) BAR(BOOST_PP_EXPAND(pred(x))) But, I think those two could cause problems with recursion. The other bug is with the preprocessors handling of whitespace. In the following example, neither concatenations work, both give '1 1' instead of '11'. #define EMPTY #define TRAILING_WHITESPACE(x) x EMPTY #define LEADING_WHITESPACE(x) EMPTY x BOOST_PP_CAT(TRAILING_WHITESPACE(1), 1) BOOST_PP_CAT(1, LEADING_WHITESPACE(1)) This causes subtle bugs with macros like this: #define BOOST_PP_LESS(x, y) BOOST_PP_IIF( \ BOOST_PP_NOT_EQUAL(x, y), \ BOOST_PP_LESS_EQUAL, 0 BOOST_PP_TUPLE_EAT_2)(x, y) This leaves a white space after the 0 when x and y are equal which prevents concatenation. This is mainly a problem with BOOST_PP_WHILE, BOOST_PP_FOLD_LEFT etc. It is possible to remove the whitespace by concatenating it to a macro which expands to nothing, but that is only possible when you know that there is definitely going to be whitespace. thanks, Daniel

3 6

RE: [boost] build_monitor.exe: a cure worse than the disease
by Aleksey Gurtovoy 03 Aug '04

03 Aug '04

Victor A. Wagner Jr. writes: > I have NO idea what criteria build_monitor uses to killing > processes, Two conditions: 1) They have to be spawned by 'bjam.exe' executable. 2) They have to be running for more than 5 minutes or show an error dialog window. > but when I awoke this morning, it had managed to kill: the dnet > "run in the background" ORG process, Pirch (my IRC client), a > proxy that lets me get dcc's through my hardware firewall, > Trillian (the instant message program), SpyBot resident..... > I think you get the picture. How did you determine that it's build_monitor that caused all that? While there is always a possibility of a program having bugs, it's hard to believe that this particular one started to misbehave all of a sudden after a year of stable work and no changes to the source code. > this thing just walked through my system like a hoard of Vandals > (or was it Visigoths). whatever I'm sorry to hear about the whole thing, but let's get to the roots of it. First of all, why do you think it was build_monitor? Second, what operating system are you running? -- Aleksey Gurtovoy MetaCommunications Engineering

2 2