December 2006 - Boost-users

[Boost.thread] Desire to check mutex locked state for debug builds
by Kai Brüning 19 Dec '06

19 Dec '06

Hi, I am new to this list, although not completely new to using Boost. When working with mutexes and locks, it is common to have functions which are part of critical sections but do not lock themselves. A typical example for this is the thread safe interface design pattern, which moves all locking to the publicly available interface functions of a class and all implementation to internal functions which expect to be called with a lock in effect. I would feel much more confident if I could check "expect to be called with a lock in effect" in debug builds. That is, if I could test whether a mutex is locked for a specific thread (in my case, "current thread" would suffice). Especially if some functions need locking and others don't and the goal is to minimize locking for performance's sake, pure convention and documentation without runtime checks seems to me dangerously error prone. Is there any way to do this test which I overlooked? Is my desire sensible? Or are there other/better ways to make sure necessary locking is done? Best regards, Kai Brüning -- RagTime GmbH Tel: [49] (2103) 9657-0 Neustraße 69 Fax: [49] (2103) 9657-96 D-40721 Hilden mailto:kai@RagTime.de Deutschland http://www.RagTime.de/

1 0

Re: [Boost-users] format: show uint8_t as a number, not a char?
by Phil Endecott 19 Dec '06

19 Dec '06

>> This does what I want: >> >> uint8_t i = 65; >> printf("i = %d\n",i); [snip] >> uint8_t i = 65; >> cout << boost::format("i = %d\n") % i; >> >> But it seems to have the same, unwanted, behaviour as cout<<i. [snip] >> At the moment I have to write things like >> >> cout << "i = " << static_cast<unsigned int>(i) << "\n";> > You can just cast the value to an int Yes, as I mentioned towards the end of my message. But I find that obfuscated. It's particularly problematic when the value is a field in a struct whose type I don't exactly remember, and I forget the need to cast it; the result is control characters being written to the terminal, which can be hard to debug. Surely, since I have asked for %d in the format string, format() can do the cast? i.e. something like (PSEUDOCODE) template <typename T> operator%(T val) { .... if (format_char='d') int val1 = static_cast<int>(val) ... } or have I missed some additional complexity? Phil.

3 2

format: show uint8_t as a number, not a char?
by Phil Endecott 18 Dec '06

18 Dec '06

Dear All, This does what I want: uint8_t i = 65; printf("i = %d\n",i); It prints "65". This doesn't: uint8_t i = 65; cout << "i = " << i << "\n"; It treats i as a char, since uint8_t is a typedef for char deep in some header file somewhere, and prints "A". I was hoping that boost::format would do what I wanted: uint8_t i = 65; cout << boost::format("i = %d\n") % i; But it seems to have the same, unwanted, behaviour as cout<<i. Is there anything that I can do to my code to get the effect that I want? If not, is there anything that can be changed in boost::format to make it do what I want? At the moment I have to write things like cout << "i = " << static_cast<unsigned int>(i) << "\n"; which is a lot of typing and obfuscates what is actually going on. Any suggestions would be much appreciated. Cheers, Phil.

2 1

vertex descriptors
by Luis Quesada 18 Dec '06

18 Dec '06

Dear all, Sorry for the newbie question! Given the index of a vertex, how can I get its descriptor? I guess function vertex (page 225 of The Boost Graph Library) does that but the function is not defined. In fact, what I want to do is to check whether an edge belongs to a graph in a context where descriptors get invalidated after vertex removals? Suppose I have done this: // create a typedef for the Graph type typedef adjacency_list<vecS, vecS, bidirectionalS> Graph; // Make convenient labels for the vertices enum { A, B, C, D, E, N }; const int num_vertices = N; const char* name = "ABCDE"; // writing out the edges in the graph typedef std::pair<int, int> Edge; Edge edge_array[] = { Edge(A,B), Edge(A,D), Edge(C,A), Edge(D,C), Edge(C,E), Edge(B,D), Edge(D,E) }; const int num_edges = sizeof(edge_array)/sizeof(edge_array[0]); // declare a graph object Graph g(num_vertices); Then I remove some vertices. Now I want to check whether edge <A,B> is still there. Calling edge(A,B,g) would not be OK because the descriptors are no longer referring to the same vertices.... Thanks in advance for your answer, Luis

6 15

[GIL] Add a constant value to an image
by Sylvain Berlemont 18 Dec '06

18 Dec '06

Hi, I would like to know what is the most readable way to write in GIL: Y = X + t, where X, Y are 2 gray32f images and t is a bits32f constant value. Best regards, Sylvain

1 0

[Announce] Mailing list outage on December 19th
by Doug Gregor 18 Dec '06

18 Dec '06

Due to a scheduled power outage at Indiana University, all of the Boost mailing lists hosted at lists.boost.org will be unavailable at the following times on December 19th, 2006: - 5:00am-9:00am Pacific US time - 6:00am-10:00am Mountain US time - 7:00am-11:00am Central US time - 8:00am-12:00pm (Noon) Indiana time - 8:00am-12:00pm (Noon) Eastern US time - 1:00pm-5:00pm GMT Doug Gregor Open Systems Lab @ Indiana University

1 0

Re: [Boost-users] BOOST_ASSERT and others
by james.jones＠firstinvestors.com 18 Dec '06

18 Dec '06

From: Sebastian Redl <sebastian.redl(a)getdesigned.at> > Maarten Nieber wrote: > > Sorry, I should have checked! > > It should be: > > > > assert( 2 + 2 == 4 && "Adding numbers failed!" ) > > That still won't fire. The concept is flawed. You cannot use assert()s > to print debug messages, because when an assert() prints a message, it > also stops the program. Why won't it fire? Let's say I normally would do this: assert( statement ); If I do this instead: assert( statement && "Statement that 2 + 2 == 4 failed" ); then if "statement" is false, the argument of assert() is false so it fires, and (more importantly), the whole argument, including the diagnostic "Statement that 2 + 2 == 4 failed" is output. Am I missing something? This seems like a clever idea to me (a bit limited, maybe, but useful for prototyping if nothing else). - James Jones Administrative Data Mgmt. Webmaster 375 Raritan Center Pkwy, Suite A Data Architect Edison, NJ 08837

1 0

How get file attributes in boost::filesystem ?
by sidle 18 Dec '06

18 Dec '06

How get file attributes in boost::filesystem ?

2 1

mpl::for_each documentation
by Jens Müller 17 Dec '06

17 Dec '06

Where can I find documentation for mpl::for_each? It's no listed in http://www.boost.org/libs/mpl/doc/refmanual/refmanual_toc.html I'm trying to understand what's going on in graph-tool's graphml-hpp in mutate_graph_impl::set_*_property ...

2 2

building boost-python under cvs (python v2.5)
by craigp 17 Dec '06

17 Dec '06

is there a way to get boost-python (cvs HEAD) built for python v2.5? i can't even seem to specify the library on the command line: >bjam --with-python-root=c:\python25 error: wrong library name 'python-root=c:\python25' in the --with-<library> option. i noticed there was a TODO for getting the python options in the jamfile... is there a work-around for now? thanks, --craig __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1