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

[BGL] Online Graph Algorithms (fwd)
by David Pearce 02 May '03

02 May '03

Hello, I am interested in how boosters think online graph algorithms should be written. I have looked at the connected_components algorithm and feel that it breaks some of the ideas behind generic programming. For example, consider the following: > typedef adjacency_list <vecS, vecS, undirectedS> Graph; > Graph G(6); > add_edge(0, 1, G); > add_edge(1, 4, G); > > // create the disjoint-sets structure, which requires > // rank and parent vertex properties > ... > initialize_incremental_components(G, ds); > incremental_components(G, ds); > > generic_algorithm_which_adds_edges(G); At this point, the connected_components data structures may be out-of-sync. Now, I suppose that we could write a wrapper class which looked like a "Graph" and provides its own add_edge to update "ds" correctly. However, this seems overly cumbersome. It strikes me that a better way of doing this would be to use an abstract base class, like so: > template<class T, ...> > class connected_components : public T { > ... > }; > > template<class T> > pair<T::edge_descriptor, bool> > add_edge(connected_components<T>::vertex_descriptor t, ...) { > // do stuff > return add_edge(static_cast<T::vertex_descriptor>(t), ...); > } Thus, we supply "Graph" as the base class and can then safely pass it to "generic_algorithm_which...". So, does anyone know of any reasons not to do this? Cheers, David J. Pearce

1 0

const shared_ptr<>
by Hossein Haeri 02 May '03

02 May '03

Dear all, I'm a newbie in boost! I want to know what the following means: int a = 5; const boost::shared_ptr<int> sp(&a); Does it mean that the following line is an error? *sp = 7; If so, what are the guarantees? Is there any documented guarantee? (I couldn't find anything in the documentation. Excuseme if I'm careless!) Thanks, --Hossein __________________________________ Do you Yahoo!? The New Yahoo! Search - Faster. Easier. Bingo. http://search.yahoo.com

3 11

RE: [Boost-Users] Is there a "known issues" page anywhere?
by Jim.Hyslop 02 May '03

02 May '03

Alisdair Meredith [mailto:yg-boost-users@gmane.org] wrote: > Is there a place where known problems for libraries/compilers are > stored? > > For instance, trying to use boost/time we hit a problem with MSVC7 and > lexical_cast unless we enable the "wchar_t as a native type" option. > While this may be good practice <g> it is not the default, > and caught us > out (and the error is none-too helpful) > > I can imagine this sort of issue is discovered and re-discovered on a > regular basis, and wonder where to look for such known issues before > using list bandwidth. There is a "Known Issues" section on the wiki (see the link in the auto-appended .sig) which is intended to document these common "gotchas". I haven't looked at it for a while, so I don't know how up-to-date it is. If that issue's not documented there, would you mind adding it? -- Jim

3 2

graph (BGL): bad documentation
by schlumpfi106 01 May '03

01 May '03

I just tried to use the boost graph library to create a spanning tree generation algorithm. After wasting a whole day I gave up and wrote all the graph code myself. I'm quiet upset about the useless online documentation for the graph library. Maybe I have to buy the book to be able to use it? Normally I would rely on some good, instructive examples and start my own program by copy-and-paste. Here, there are only a few examples and they are either too simple or too specific. Most documentation topics provide no examples at all. Sometimes I head to guess through the template-loaded headerfiles because there was no documentation. I failed in creating a graph from a given adjacency matrix. I failed in understanding the "interior property" concept. I failed in printing a graph and its properties to graphviz. To the developers: Please, before you implement anything new -- document the old things.

1 0

RE: [Boost-Users] Regexp Homework :-)
by Ben Hutchings 01 May '03

01 May '03

Simon Brown <simon.brown(a)kns.ch> wrote: > A few days ago I received great help with a regexp question - > now here's > another which I doubt can be done. > > Question: if the text string contains three words AAA BBB AAA > then return > the string (if any) which is repeated two or three times. > > Examples: > AAA AAA AAA return AAA > AAA BBB AAA return AAA > AAA AAA BBB return AAA > AAA BBB CCC return <nothing> > > Can this be done with a regular expression? Match against "(\\w+) (?:\\1 \\w+|\\w+ \\1)|\\w+ (\\w+) \\2" and return the match for whichever sub-expression (1 or 2) was matched. (I tested this with Python not Boost regex, but they should be compatible.)

2 1

Is there a "known issues" page anywhere?
by Alisdair Meredith 01 May '03

01 May '03

Is there a place where known problems for libraries/compilers are stored? For instance, trying to use boost/time we hit a problem with MSVC7 and lexical_cast unless we enable the "wchar_t as a native type" option. While this may be good practice <g> it is not the default, and caught us out (and the error is none-too helpful) I can imagine this sort of issue is discovered and re-discovered on a regular basis, and wonder where to look for such known issues before using list bandwidth. -- AlisdairM

1 0

regexp match question
by Simon Brown 01 May '03

01 May '03

Hi All, Here's a question I can't answer. I'm using regexp to dig out strings from data received via Ham Radio, so the txt often is corrupted. What I want to match is two consecutive callsigns separated by whitespace, for example 'hb9drv hb9drv' but not 'hb9drv hb9drX'. I can define the match condition for a callsign, but can I use the match option {2,3} somehow to match the same callsign two or three times? Effectively match the same word two or three times? Help received with much thanks, Simon Brown, Casa Bergenia, CH-7031 Laax www.hb9drv.ch [Non-text portions of this message have been removed]

3 5

Licence terms
by Ben Hutchings 30 Apr '03

30 Apr '03

Where should I post to ask questions and make suggestions about licence terms for Boost libraries? I am specifically concerned about the large number of slightly different wordings of some of the licences, which I think would all have to be included in documentation of a program using the covered libraries.

1 0

RE: [Boost-Users] const shared_ptr<>
by Tom Matelich 30 Apr '03

30 Apr '03

> From: Hossein Haeri [mailto:powerprogman@yahoo.com] > Sent: Wednesday, April 30, 2003 8:29 AM > > I want to know what the following means: > > int a = 5; > const boost::shared_ptr<int> sp(&a); > I assume you know this, but this code is in error. When the last shared_ptr referencing &a goes away, it will try to delete a, which is illegal. There is a workaround defined at http://www.boost.org/libs/smart_ptr/sp_techniques.html#static , though. ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

1 0

ptime and time_duration
by Andre du Toit 30 Apr '03

30 Apr '03

Hi. Since time_duration uses <long> to represent each of the fields (H:M:S:FS), I should be able to add any hour offset to a ptime, as long as the resultant ptime is inside the date/time range representable by boost? With the date ranges of boost being 1400-Jan-1 00:00:00 -> 10000-Jan-1 00:00:00, the number of hours between these dates is (very roughly!) 7.534E7. This is less than the <long> limit on my system, which is +/- 2.147E9. That is good. So; doing the following: ptime t_max( date(10000,Jan,1) ); ptime t_min( date(1400,Jan,1) ); time_duration td = t_max - t_min; std::cout << to_simple_string( td ) << std::endl; should produce something like : 7.534E7:0:0 or if you want to write it out: 7533600:0:0 not? Instead; the output I get is : 224112:19:12. This corresponds to a timespan of roughly 25 years which means I just ran into some overflow/underflow problem somewhere. What gives? What time range is supported by time_duration if it does not use the <long> limits? I have difficulty deducing this information from the header files, the source for the lib is not too helpfull either! What am I missing here? Greetings, Andre.

2 1