- Boost - lists.boost.org

boost [msm] deferring event by base type.
by Richard Szabo 25 Jun '10

25 Jun '10

I try to use the following code struct base {}; struct event1 : public base {}; struct MyState1 : public msm::front::state<> { typedef mpl::vector<base> deferred_events; }; than MySm.process_event(event1()); when state machine in MyState1 causes that event gets dispatched and not deferred the same code works if I replace typedef mpl::vector<base> deferred_events; with typedef mpl::vector<event1> deferred_events; than it works fine and event1 is deferred in MyState1; It would be nice to able to defer event by it's base types it would make code maintenance easier not have to declare all events one by one for deferring. Br. Richie

2 5

[filesystem][v3] Class path codecvt argument now available for appropriate functions
by Beman Dawes 24 Jun '10

24 Jun '10

The initial V3 implementation, like V2, could alter the codecvt facet used for conversion via a static imbue() function. This has a number of problems, including: * It is, in effect, a global variable, and thus "considered harmful" as being prone to abuse. * It is subject to data races in multi-threaded environments. * It can be hard to use. * It can be hard to explain. As of revision 63283 this morning, V3 class path in trunk now has an additional "const codecvt&" for all functions where conversion may be required. So you can now write: utf8_codecvt_facet utf8; // UTF-8 facet from boost or other sources path p(some_string, utf8); // some_string will be treated as UTF-8 encoded ... operations on p That's great, but where does that leave path::imbue? I think it should be deprecated. What do others think? --Beman

3 3

[asio] Re: Pass some parameter to io_service::run
by Dmytro Ovdiienko 24 Jun '10

24 Jun '10

James, it is not academic reason. I need latency 1-2us. Right now I have 15us. I cannot waste time on locks and I cannot create bottlenecks. And how about just let user pass "void *"? I know it is not type safe... maybe boost has something like type-safe variable argument list? On 11 January 2010 17:13, James Mansion <james(a)mansionfamily.plus.com>wrote: > Dmytro Ovdiienko wrote: > >> >> I don't understand why you would have an objection to a permanent >> binding with TLS >> >> TLS looks like global variables. Global variables violate encapsulation. I >> suppose to support my program at least 2 years. I don't want to hate myself >> two years :) >> > I think that's quite a silly reason. So are singletons, but they are both > useful. ... > > >> or >> a temporary binding using a pool of parsers, since it seems that >> each datagram must be >> parsed entirely and then the parser reset for the next? >> >> Exactly. But I don't want to create pool of the parsers. Access to the >> pool should be synchronized. Synchronization eats CPU times. >> > ... but not nearly as silly as this one. Getting an item from a pool is a > great candidate for a spinlock and if you're concerned you can easily > segment the pool. Its a non-issue. I defy you to measure it in the context > of all teh context switching and system calls you're doing. > > Of course, the pool is a singleton. And that's like a global. > > The right tools to get you what you want are right there - you're just > choosing not to use them for academic reasons. > > James > > -- Dmytro Ovdiienko e-mail: dmitriy.ovdienko(a)gmail.com skype: dmitriy.ovdienko(a)gmail.com mobile: +38050-1909731 -- Dmytro Ovdiienko e-mail: dmitriy.ovdienko(a)gmail.com skype: dmitriy.ovdienko(a)gmail.com mobile: +38050-1909731

1 0

[thread] once_flag causes dynamic initialization?
by Andrey Semashev 24 Jun '10

24 Jun '10

Hi, I've noticed that on Windows using call_once is not thread-safe as the once_flag structure has a destructor which makes it dynamically initialized. I can see the ticket [1] for this problem is hanging for quite a while, whereas the bug is quite critical, IMHO. Are there any plans on fixing it, perhaps by reverting the change that introduced it? [1] <https://svn.boost.org/trac/boost/ticket/4225> PS: Curiously, the change was made on April, 1st. A long-term joke? :)

1 0

[standard] mutable references
by Gennady Proskurin 24 Jun '10

24 Jun '10

According to 7.1.1-10, reference members cannot be mutable. I think its worth to apply attached patch. ---

1 0

[chrono] Documentation suggestions
by Roland Bock 24 Jun '10

24 Jun '10

Hi, I have a few remarks/suggestions for the documentation of Boost.Chrono: *General: *http://svn.boost.org/svn/boost/sandbox/chrono/libs/chrono/doc/html/index.html There are several sections which attract my attention when trying to get a grip on the library: * Getting Started * Tutorial * Examples Each of those looks a bit like a random collection of valuable hints to me (sorry, if I missed the pattern). My personal troubles in understanding the basics are documented in the mail referenced in the tutorial section. I think it would be worth it to reorganize all three into one section (e.g. tutorial) which shows how to utilize the different parts of the library. For instance with the following subsections: * Clocks (including stopwatch and thread_clock) * Formatters * Reporters * Conversions (see also Tutorial section below) * ...? Each section would contain usage patterns and complete sample programs. *Users Guide: *http://svn.boost.org/svn/boost/sandbox/chrono/libs/chrono/doc/html/boost_chrono/users_guide.html * Personally, I would move the references section to the appendices (I clicked there several times when I really wanted to see the Reference section) *Reference: *http://svn.boost.org/svn/boost/sandbox/chrono/libs/chrono/doc/html/boost_chrono/reference.html * thread_clock is missing in the reference section * typo "chorno.hpp>" on several pages *Examples:* http://svn.boost.org/svn/boost/sandbox/chrono/libs/chrono/doc/html/boost_ch… * There are several hints like "See the file i_dont_like_the_default_duration_behavior.cpp". It would be nice to have a working link to the file instead * Some examples are not linked to a file. I think it would be good to have a working sample file for each example *Tutorial:* http://svn.boost.org/svn/boost/sandbox/chrono/libs/chrono/doc/html/boost_ch… * The section "How to define a thread clock" is now obsolete, isn't it? thread_clock is now integral part of chrono. * An example on how to use the thread_clock would be very helpful. * An example on how to obtain a duration in some unit would be helpful See here (http://lists.boost.org/boost-users/2010/04/58204.php) for what I meant with the last item. The answers by Howard Hinnant in this mail were extremely helpful and could probably be re-used in the tutorial. Regards, Roland

2 2

Re: [boost] [filesystem] wfstream file open does not create file in std::ios_base::out mode
by Ramakrishna, Kirankumar IN BLR SISL 23 Jun '10

23 Jun '10

Hi, I upgraded to boost version 1.43, but that did not help. Still file creation fails with wfstream and error code set is ERROR_FILE_NOT_FOUND. Below code captures the error codes #include "boost/filesystem/fstream.hpp" int main() { boost::filesystem::path pA("C:\\testA.txt"); boost::filesystem::wpath pW(L"C:\\testW.txt"); boost::filesystem::fstream fileHandleA; boost::filesystem::wfstream fileHandleW; fileHandleA.open(pA, std::ios_base::out); cout << "fileHandleA.fail() return value is: " << fileHandleA.fail() << endl; cout << "Error code is: " << GetLastError() << endl; fileHandleW.open(pW, std::ios_base::out); cout << "fileHandleW.fail() return value is: " << fileHandleW.fail() << endl; cout << "Error code is: " << GetLastError() << endl; return 0; } Am I doing something wrong? Regards, Kiran On 2010-06-19 00:44, Beman Dawes wrote: > On Thu, Jun 17, 2010 at 9:48 AM, Ramakrishna, Kirankumar IN BLR SISL > <Kiran.Ramakrishna@???> wrote: > > Hello, > > > > I am using boost version 1.42 and the below code is tested on Windows Xp SP3 and Suse 11.2 > > > > The below code creates new file when opened using fstream object, but fails to create when opened using wftream object. > > This behavior is consistent across platforms (Windows and Linux). > > Also the error code which I get after fileHandleW.open call is, > > > > > > Â 1. Â ERROR_FILE_NOT_FOUND for Â std::ios_base::out > > Â 2. Â ERROR_NEGATIVE_SEEK for std::ios_base::app > > > > In case of append mode, the file gets created. But still I get the above mentioned error code. > > > > Important notice: This e-mail and any attachment there to contains corporate proprietary information. If you have received it by mistake, please notify us immediately by reply e-mail and delete this e-mail and its attachments from your system. Thank You.

2 3

Question about exposing template class to python using boost::python
by Yongning Zhu 23 Jun '10

23 Jun '10

Hi, I implemented a template<class T, int d> VECTOR; Now I can expose specialized class like VECTOR<int,2> to python, but I want to expose more specializations without copying a whole set. What shall I do? I found some early posts about this problem, but didn't figure out a solution. Is class_builder the one I should look for? Thanks in advance? Yong

3 3

Re: [boost] [GIL] kth_channel_view
by fabien.castan＠free.fr 23 Jun '10

23 Jun '10

>> #include <boost/gil/gil_all.hpp> >> ... >> kernel_1d<float> kernel2(gaussian_1,9,4); >> convolve_rows<rgb32f_pixel_t>(kth_channel_view<0>(const_view(img)),kernel2,kth_channel_view<0>(view(img))); >> jpeg_write_view("out-convolution2.jpg", view(img)); >> ... >> >> This doesn't compile. > >Does it compile when using view(img) instead of const_view(img)? >Please provide a complete minimal code sample displaying your problem. >Makes my life easier. ;-) I'm really sorry, I just notice that it's a stupid mistake from me... I forgot to change the template parameter of the concolve_rows function ! If I use k_channel_view I manipulate a grayscale image... kernel_1d<float> kernel2(gaussian_1,9,4); convolve_rows<gray32f_pixel_t>(kth_channel_view<0>(const_view(img)),kernel2,kth_channel_view<0>(view(img))); jpeg_write_view("out-convolution2.jpg", view(img));

2 1

[GIL] kth_channel_view
by fabien.castan＠free.fr 23 Jun '10

23 Jun '10

Hi, I have some difficulties using kth_channel_view. I'm trying to modify the "convolution.cpp" example, like that: #include <boost/gil/gil_all.hpp> ... kernel_1d<float> kernel2(gaussian_1,9,4); convolve_rows<rgb32f_pixel_t>(kth_channel_view<0>(const_view(img)),kernel2,kth_channel_view<0>(view(img))); jpeg_write_view("out-convolution2.jpg", view(img)); ... This doesn't compile. Regards, Fabien

2 1