- Boost-users - lists.boost.org

Boost 1.30 with g++ 3.4
by Ben Hutchings 02 Oct '04

02 Oct '04

I have an existing body of code using some parts of Boost 1.30 (Bind, Function, Regex, Thread, and some utilities) and compiled using VC++ 7.1 on Windows. I now need to build with g++ 3.3 or 3.4 on Linux as well, and for various reasons I would prefer to use 3.4. Boost 1.30 predates g++ 3.4 by some way and I know that it won't all work without changes. However, neither will 1.31, and 1.32 has still not been released. I would rather not upgrade to either of these in case they break the existing Windows build. Are there any known bugs or other problems in the above-mentioned libraries from Boost 1.30 when used with g++ 3.4, and if so, any known fixes for them? (Assume I'm using Boost 1.30.2 - I might still have 1.30.0 but upgrading to 1.30.2 should be low-risk.) Ben.

2 1

Are BOOST_BUILD_PATH && BOOST_ROOT necessary?
by Steven T. Hatton 01 Oct '04

01 Oct '04

>From reading the documentation on boost.org, I have gotten the impression I don't need to set $BOOST_BUILD_PATH and $BOOST_ROOT. I'm more than a little confused about how I should use the boost libraries. Part of this comes from the fact that it is not clear to me exactly where the demarkation is between using the build system to build boost, and using it to build my own programs. Some of my confusion comes from the appearant change in configuration requirements described in the documentation. It is suggested I don't need the environment variable used with earlier releases. One of these variables is $BOOST_ROOT, which the documentation used as an example of how to build boost. That example was exactly what I successfully (I believe) build the libraries with. I'm using SuSE Linux 9.1 on a P4 box. Some basic questions are these: When I downloaded and built boost, I first fetched the source for bjam. This is the name of the file I downloaded: boost-jam-3.1.10.tgz It compiled quite nicely, and I copied the resulting bjam jam mkjambase and yyacc to a location in my path. Are there anything other things such as configuration files I need to reside with the binary executables? Are there necessary configuration file that I need to have visible to these programs, even if they are not colocated? I then build boost with the command: bjam --prefix=$BOOST_ROOT "-sTOOLS=gcc" install after setting my BOOST_ROOT the way I wanted it. That resulted in two directories being placed in $BOOST_ROOT/ one is called $BOOST_ROOT/include/ the other is called $BOOST_ROOT/lib/. Under $BOOST_ROOT/include/ I have <path>/boost-1_31/boost/ which contains dozens of header files. I added $BOOST_ROOT/include/boost-1_31 to my $INCLUDE_PATH, and $BOOST_ROOT/lib to my $LIBRARY_PATH and $LD_LIBRARY_PATH variables. Is this the correct way to configure my system to use boost? Is there anything else I need to do in order to compile my programs against the boost headers and libraries? Will the boost libraries work as well as other, comperable libraries using autotools rather than bjam? Are there any special accommodations I need to make in order to use boost with the gnu build tools? Do I need to do anything special to use boost with bjam that does not involve placing files in my current project directories? For example, do I need to provide a global configuration file of some kind? Is it a preferred option, even if not strictly required? There is much remaining in the source directory where I compiled boost. Is that something I need to also put in an accessible place on my system? If so, what so I need to put where, and how to I modify my environment variable to facilitate its use? -- Regards, Steven

4 15

BOOST_CHECK_CLOSE apparent misbehavior
by Ross Boylan 01 Oct '04

01 Oct '04

I get the following error Specification_test.cc(22): error in "specsimple": difference between spec(i, j){0.206788} and results[k]{0.206788} exceeds 1e-05% The numbers agree to all printed digits. I get a maximum relative error (assuming the printed values are rounded from extremes)*(see footnote) of .000001/.206787 = 4.8e-6 < 1e-5. Even doubling this is under the tolerance. As I read the documentation, this should pass. 1e-6 tolerance produces a lot more discrepancies; 1e-4 says they are all OK. Have I misunderstood something, or is something funny going on? In particular, the "%" after "1e-5" in the error message suggests the tolerances are supposed to be in percentage terms, i.e, 1e-5 implies a real tolerance of 1e-7. That would explain things, but I see nothing in the documentation to suggest the tolerance value is a percent. (Well, the docs do say the tolerance can be an integer.) The relevant line is BOOST_CHECK_CLOSE(spec(i, j), results[k], tol); All relevant quantities are doubles. It may be relevant that the program under test includes <cmath>. Debian package libboost-dev 1.31.0-7 on an Athlon system. # g++ --version g++.real (GCC) 3.3.4 (Debian 1:3.3.4-6sarge1) Linux kernel 2.4.26 * In fact, the results value is coded as 0.2067880 in the program, so my calculation likely gives double the possible discrepancy.

2 1

RE: [Boost-users] Re: Boost:regex and C++ Parsing - Spirit!
by Foster, Gareth 01 Oct '04

01 Oct '04

Sheesh, I've learned some new words there. I mean, dictionary.com is as close to naked ... Only joking. I thought it was the perfect way to describe how slick the library was :) that's my effusing done for the day anyway. -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Justin Findlay Sent: 30 September 2004 17:22 To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Re: Boost:regex and C++ Parsing - Spirit! On Thu, Sep 30, 2004 at 10:23:22AM +0100, Foster, Gareth wrote: > > Wow, I just read through the 1st few pages of the spirit docs, and I > can safely say, that is as close to naked women as C++ code will ever > be, its such a sexy idea. One library, and you can parse any sort of > config file, code or anything! Amazing. Gentlemen, consider decency. /me covers his eyes in embarassment. > Interetsing stuff, thank you all! Spirit is indeed an elegant library, by utilizing many powerful features of C++ in context it exemplifies what sublime things can be done with the language. Good luck with your project Gaz. (-: Please excuse the offtopic post. No effusions of remonstrance were personally directed. Justin _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Forgot to add
by Foster, Gareth 01 Oct '04

01 Oct '04

I also was going to say Joel, you ought to moan about the description of spirit in the docs, I can see it is there in the list of libraries now, and whilst it probably does describe what the library does, it doesn't do it in a way that is very identifiable for those who don't know about the subject, I just plain missed it, despite the fact that I was looking to parse C++. spirit <file:///C:/boost/boost_1_31_0/libs/spirit/index.html> - LL parser framework represents parsers directly as EBNF grammars in inlined C++, from Joel de Guzman and team. Why not something like ... spirit <file:///C:/boost/boost_1_31_0/libs/spirit/index.html> - A customisable framework for parsing any syntax. You will know what I mean and how to describe it properly etc I'm sure. Just a thought :) Gaz

2 1

Boost:regex and C++ Parsing - Clarification
by Foster, Gareth 30 Sep '04

30 Sep '04

Hello, Thank you Caleb and Hartmut for your replies. You both seem to think regex is a bad way to go, I will explain better what I want to write just to be clear. I want to write a tool (cli probably) where I can say, here you go, here is a large folder full of code, go and parse it. I will store the results in XML format somewhere, then, I can do say, "<myapp> class someclass" and the program will go and find where that class is declared/defined using its database, saving me headache. So I thought I could use one of the C++ expat wrappers, and boost regex looked powerful enough to do the parsing if only I were handy enough with regular expression syntax. Anyway, I don't know if that better explanation will make any difference to you recommendations, I look forward to reading you opinions. Oh, and the example I looked at is here Hartmut: http://boost.org/libs/regex/example/snippets/regex_search_example.cpp - that is what got me thinking I might actually be able to take on this challenge. Okay lads, thanks again, cheers Gaz -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Hartmut Kaiser Sent: 29 September 2004 18:41 To: boost-users(a)lists.boost.org Subject: RE: [Boost-users] Boost:regex and C++ Parsing Foster, Gareth wrote: > I have a couple of questions, firstly, how might I extend the > example for parsing C++ code for class names so that it > records the line number on which the class is defined? I > thought maybe I could extend the regular expression so that > it has "|(\n)" at the end, or maybe there is another way. In > any case I am not sure if that is the correct way to extend > the regex and I am unsure how to check the regex_match result > to see if it was a new line character I encountered or a class name. Which example you're referring to? > Secondly, are there any efforts anywhere to parse C++ for > other keywords by this approach? I don't know of any efforts going on regarding C++ parsing with the help of regex (non-authoritative answer). But there is the Wave library (Boost review is due shortly), which is a C/C++ preprocessor containing different C++ lexing components, which may be helpful for you during writing a class name extraction tool. Regards Hartmut _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

3 2

RE: [Boost-users] C++ Parsing - oops
by Foster, Gareth 30 Sep '04

30 Sep '04

Hey mate, I've actually come across this, whilst it is really neat, it does require include paths, infact, probably a whole make system to resolve names etc, which I won't always have, so I decided against using it. If I'm wrong on that, please do let me know, since it would be ideal other than that one issue. Thanks! Gaz -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Volker Boerchers Sent: 30 September 2004 15:30 To: 'boost-users(a)lists.boost.org' Subject: Re: [Boost-users] C++ Parsing - oops Hi Gareth, > So, will spirit be the best way then? Glad to hear its already a part of > boost btw, congratulations mate. there might be another option. The gcc-xml project (http://www.gccxml.org) lets you use gcc for parsing. Volker -- Volker Börchers TECON Systems AG, http://www.tecon.de _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Boost uBLAS
by Pascal Bauer 30 Sep '04

30 Sep '04

Hi Boost People, Have just a simple Question: 1. I use the uBLAS matrix class together with the MPI. Is it save to send matrices similar to std::vectors, by using &(Matrix(0, 0)) as the argument for the sending buffer? (Similar as one can do &(Vector[0]) for the std::vector? ) I just use matrix<double> in the code, without any further template specifications. Thanx P.

1 0

Boost.graph read_graphviz problems in CVS head
by Jeffrey Holle 29 Sep '04

29 Sep '04

I attempted to switch from boost_1_31_0 to the CVS head available on Sun Sep 26 20:25:26 PDT 2004. My application uses read_graphviz to read the attached file. When built with the CVS head boost and run, I get the message "parse error" and a segment fault. When build with boost 1_31_0 and run, it works fine. digraph Components { node [ shape = record ]; size="6,6"; 0[label="0",type="Component"]; 1[label="1",type="Component"]; 2[label="2",type="Component"]; 3[label="3",type="Component"]; 4[label="4",type="Component"]; 5[label="5",type="Component"]; 0 -> 1[type="Association"]; 1 -> 4[type="Association"]; 4 -> 0[type="Association"]; 2 -> 5[type="Association"]; }

2 1

Boost:regex and C++ Parsing
by Foster, Gareth 29 Sep '04

29 Sep '04

Hello, I have a couple of questions, firstly, how might I extend the example for parsing C++ code for class names so that it records the line number on which the class is defined? I thought maybe I could extend the regular expression so that it has "|(\n)" at the end, or maybe there is another way. In any case I am not sure if that is the correct way to extend the regex and I am unsure how to check the regex_match result to see if it was a new line character I encountered or a class name. Secondly, are there any efforts anywhere to parse C++ for other keywords by this approach? Many thanks for any help or hints on this, for the record I am really impressed with boost, a while ago, I might have said I didn't want the depandancy, but now I'd put up with that any time, good luck getting boost work into the next C++ standard, it's great! Regards Gaz

3 2