August 2005 - Boost-users - lists.preview.boost.org

Regexp without ^ and $
by Merrill Cornish 02 Aug '05

02 Aug '05

I am using regex_match() from Boost 1.32.0. I am well famililar with Perl's regular expression, but I got a surprise from Boost's regex_match(). My basic code is static const regex CHAR(" char "); regex_match(newSpec, CHAR) If newSpec is " char ", then it matches, of course. However, if newSpec is " char * ", it doesn't match. Hmmmm. In my Perl experience, a regex of " char " says match those six characters if found ANYWHERE in newSpec. This would mean that a Boost regex of " char " would match a newSpec of " char * ", but it doesn't seem to. If I change the regex to ".* char .*", then it matches, implying that Boost's regex of " char " is being treated as if it were Perl's "^ char $". I could add the ".*" before and after each of my (many) regular exressions. However, I'm guessing there is a match_flag_type value which would do this for me. Unfortunately, none of the match constant descriptions seem to be what I need. How should I write a Boost regex so that it would find " char " anywhere in the input string regardless of any prefix or suffix? Merrill

3 2

a few questions about boost:any
by Shishir Ramam 01 Aug '05

01 Aug '05

Hello fellow boosters! I am looking to use boost:any. While gaining an understanding of how to use it, it strikes me that the real purpose it serves is that the virtual table functionality is exposed to the user. It seems that efficiency of table lookup are then the application responsibility. My first question - Is this a reasonable understanding? I have been looking to see if tag dispatching can somehow circumvent the cost of table lookup that would be necessary, and the only thing that strikes me is to - store a type id within the type boost::any represents, - retrieve this id and either use it in an if/else structure, or a function pointer lookup in an array. A complete tag dispatch implementation seems impossible, since the type id within the boost::any structure is a run time quantity. My second question - Are there better ways to do it than an index into an array of functions? TIA. -shishir -- It is not because things are difficult that we do not dare, it is because we do not dare that things are difficult. - Seneca

1 0

boost python + vc 7.1
by Uwe Schmitt 01 Aug '05

01 Aug '05

Hi, I have a problem with compiling an application using boost python. I use Microsofts VC++ version 7.1. The following lines describe my problem: cl /EHsc /TP /IC:\Python24\include /ID:\home\Administrator\svn_local\public.procoders.net\pcsvm /IC:\boost\include\boost-1_32 /c pythonAdapter.cpp /FopythonAdapter.obj pythonAdapter.cpp C:\Boost\include\boost-1_32\boost\python\operators.hpp(209) : fatal error C1001: INTERNER COMPILERFEHLER (Compilerdatei 'msc1.cpp', Zeile 2701) Whlen Sie im Men '?' von Visual C++ den Befehl 'Software Service', oder ffnen Sie die Hilfedatei fr den Software Service, um weitere Informationen zu erhalten Any hints what to do ? Greetings, Uwe

2 1

Include boost in software distribution?
by Joseph Turian 01 Aug '05

01 Aug '05

Hi, I'm packaging some software for distribution, and the software depends upon the 'program_options' library. I don't want to assume that the user has boost installed, or ever jam. So I was wondering about the potential for automagically building and using the 'program_options' library if it is not already present on the user's system. I would include the source files in libs/program_options/, and write a small Makefile for it. I would also include all the Boost headers. Is this sufficient, even though it wouldn't run the Boost configure script to determine system-specific settings? Can you see any problems in building this library without doing a full Boost install? Thanks! Joseph -- http://www.cs.nyu.edu/~turian/

4 5