June 2009 - Boost - lists.boost.org

Re: [boost] Proposal: Monotonic Containers
by Artyom 10 Jun '09

10 Jun '09

> The space has already been created before the allocator is > requested to > provide for new storage. > > Otherwise, we could never use a vector<char> on a > machine that requires > 16-byte alignment. > This is not correct. Allocators allocate memory as single chunk. So, if you use vector<char> that has size of 10 characters, it would allocate a single chunk of at least 10 bytes, and it's beginning would be aligned to 8 or 16. So if you for example, allocate 3 charrecters that start at 0x100 to 0x102, and then you would try to allocate single integer you need to return pointer with value 0x104, 0x108 or 0x110-- aligned to 4, 8 or 16 according to architecture requirements, if you would return 0x103 it may cause illegal operation on some architectures like ARM, hundreds cycles cost on architectures like IA64 and fault in execution of atomic operations like in shared_ptr. You **must** return aligned pointers from allocators. Artyom

3 4

Re: [boost] [multi_array] Negative strides
by pete＠pcbartlett.com 10 Jun '09

10 Jun '09

>John Reid <j.reid(a)mail.cryst.bbk.ac.uk> wrote:John Reid wrote: >> Either the library was designed with negative strides in mind and it is >> just the assertions that are stopping me using them or they were never >> part of the original spec and are not safe to use. In either case I >> can't find them mentioned in the documentation. Could the owner of the >> library clear this up? >Do libraries have owners? Posts on boost.multi_array don't seem to >provoke many responses. Nearly all libraries have owners (see e.g. http://gitorious.org/boost/svn/blobs/master/libs/maintainers.txt ). There is considerable variation in how closely each owner is monitoring this list. Pete

1 0

Re: [boost] [Christian S] Spirit and ANTLR - Request for example
by Christophe Henry 10 Jun '09

10 Jun '09

> brought an idea to boost and had it shot down [...] I quite agree with Luke's explanation and while Christian's behaviour is not perfect, I'd like to point out that bringing a new idea to boost for the first time is a stressful experience, as I myself remember full well. And having his idea immediately shot down is even less nice (as I also remember full well). It seems to happen quite sometimes that people coming with new ideas are met with (unconsciously) not so friendly behaviour. Perhaps we should think about the way new people are met if we don't want to end as a closed (and thus condemned) community... And in the same time, maybe Christian could take a short break and we could all try to restart on a better basis. I therefore offer to close this thread for now as it clearly brings nothing but spam in our mailboxes. If Christian comes up later with interesting examples, we would all profit from them. Regards, Christophe

2 1

Re: [boost] Proposal: Monotonic Containers
by Artyom 10 Jun '09

10 Jun '09

Hi, Several notes: 1st... This code does not compile with gcc-4.3 (mingw) a) Fix: map.h/set.h --> - set(Predicate P, Allocator const &A) - : Set(P, A) + set(Predicate Pr, Allocator const &A) + : Set(Pr, A) because it shadows template parameter P. b) Fix in ./boost/monotonic/storage_base.h Add #include <boost/cstdint.hpp> 2nd... Never use: #pragma once not portable 3rd... Your allocator is not correct you do not align the pointers as required by most of architectures. You need to provide 8-16 bytes alignment... According to system 4th... Why don't you use boost::pool? It is debugged and tested instead of writing new implementation? 5th... Implement ALL constructors of each container. For example I can call. std::string str="Foo"; std::vector<int> v(str.begin(),str.end()); this is not supported 6th... This is bad idea to derive from std containers. They to do not provide virtual destructors. This may be very error prone. This is just short glance. Artyom --- On Tue, 6/9/09, Christian Schladetsch <christian.schladetsch(a)gmail.com> wrote: > From: Christian Schladetsch <christian.schladetsch(a)gmail.com> > Subject: Re: [boost] Proposal: Monotonic Containers > To: boost(a)lists.boost.org > Date: Tuesday, June 9, 2009, 3:06 PM > Hello, > > That is not what I asked for. I wanted online docs and > code. We are very > >> busy people, and anything that takes extra time is > likely to mean we don't > >> bother. > > > > > OK well on second thought, what? > > I am also busy. > > I gave you a link to code and documentation in the one > archive. > http://www.boostpro.com/vault/index.php?action=downloadfile&filename=Monoto… > > Click on the link, get the code and the documentation. > Sorry that you have > to click into the archive twice, but boost doesn't have a > better way of > doing it ATM that I am aware of. > > Chrstian. > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost >

2 2

Re: [boost] identifying a string type
by Steven Ross 10 Jun '09

10 Jun '09

Is there a boolean argument that can be used to identify whether a data type is some form of string that has a [] operator that returns an integer data type (preferably 2 or less bytes in size)? I now am using enable_if so that strings and floats are identified and the proper specialized call is made for them when spread_sort is called by the user. It would be nice to automatically identify strings too inside enable_if, so that the user can just call "spread_sort" and the correct specialized and optimized sorting call will be made depending on the data type for integers, floats, and strings, and std::sort will be called for unsupported data types.

5 7

Proposal: Extended Allocators (was Montonic Containers)
by Christopher Jefferson 10 Jun '09

10 Jun '09

It is my belief (but I'm happy to shown wrong) that the idea of 'Monotonic Containers' to provide a new allocator for vector to support a small block of memory is fundamentally not going to work, due to 2 limitations of allocators: 1) There is no support for in-line expansion (effectively, realloc which doesn't move if it can't realloc) 2) There is no support for an allocator to say "No, you can't have X memory, but I could give you Y", so it's hard to "exactly fill" a limited memory buffer. I have been working around this by implementing new special containers, which know about the size of a fixed buffer and therefore can make use of the space. I now think I'm probably going about this the wrong way, a much better way would be to implement: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n2045.html , or a variant thereof. The big problem with doing this "within boost" is that it would pull in a basically complete implementation of std::allocator and the containers into boost. In an ideal world, such a small and in my opinion useful extension would have been added to C++0x, but it's far too late for that now. tl;dr : Seeing as we are having a "what would boost accept" couple of days, would adding an (extended) implementation of a substantial part of the standard library be acceptable, or would I be better off trying to get this picked up by one of the existing standard libraries to pick this up (my worry being they may not want, quite sensibly, want to accept code which is not standardised). Chris

3 2

boost::directx?
by Christian Schladetsch 09 Jun '09

09 Jun '09

Hello, Is there a possibility of making submissions that use meta-C++ to assist with writing DirectX-based code? Obviously, DirectX is Microsoft-specific (but not necessarily platform specific!). So my question comes down to "is there a space in boost for MS-specific systems"? Perhaps in boost::microsoft or boost::directx or boost::microsoft::directx. I have a number of systems that are generally useful, and would like to support the use of boost in game development. It is hard enough to get gamedevs to use STL, let alone boost. But for example, creating vertex declarations is a royal PITA and can be assisted with some deft use of boost::mpl. Plus, we are already using boost::shared_ptr<> to make com_ptr<>, but it could be made more obvious and a lot of people don't know how to use it. There are two ways of looking at it. First, boost is about C++ and not about supporting specific API's. Fair enough - but then there are many boost libraries that use system-specific idioms such as threads and lock-free containers. It could be argued that boost doesn't want to tie itself to a proprietary API which itself will be versioned and modified externally to boost, but the same is already true in a lot of what boost does. Another way of looking at it is that boost is here to be used by real people solving real problems. In this case, it seems that there is room within a new namespace for domain-specific boost libraries. I realise this is a can of worms. What do you think? Should boost be more egalitarian and increase its possible audiance, or is introducing a dependancy on a 3rd party proprietary API an anathema to boost? Regards, Christian. ----- http://www.linkedin.com/in/christianschladetsch

13 40

[Christian S] Spirit and ANTLR - Request for example
by John Phillips 09 Jun '09

09 Jun '09

Christian, Inside the directx thread you asserted that there are things that ANTLR can do and Spirit can't. This met with some resistance, but I would like to take a different approach. Please post a working code example, along with rationale for why you want to do it, that shows something ANTLR can do, but Spirit can't. The goal is to move this out of the land of unsupported allegations, and into specific examples that can promote specific responses. My experience of Joel and Hartmut suggests that if you have any such real examples, they will use them as a basis for improving Spirit, a process which benefits everyone. If you display examples, and Hartmut and Joel (or anyone else) respond by providing working examples of how Spirit does that, too, then better information is propagated on the list, and again everyone benefits. In general, this is a better way to critique by comparison. Provide actual working code for the comparison, and there is far less room for misinterpretation of intent or conclusions. John

14 27

Borland tests
by Robert Ramey 09 Jun '09

09 Jun '09

The serialization library fails to build on the latest borland compiler with the error message quoted below. This suggests to me that some changes might be in order for config/borland and/or perhaps iterator_facade.hpp Robert Ramey "C:/BuildAgent/work/e5a72339587b6bb4/cb2009_3/bin/bcc32" -j5 -g255 -q -c -P -Ve -Vx -a8 -b- -v -Od -IC:\BuildAgent\work\e5a72339587b6bb4\spirit -tWC -tWR -tWC -tWD -WM- -DBOOST_ALL_NO_LIB=1 -DBOOST_SERIALIZATION_DYN_LINK=1 -I".." -I"C:/BuildAgent/work/e5a72339587b6bb4/cb2009_3/include/" -o"C:\BuildAgent\work\e5a72339587b6bb4\results\boost\bin.v2\libs\serialization\build\borland-6.1.3\debug\basic_text_wiprimitive.obj" "..\libs\serialization\src\basic_text_wiprimitive.cpp" ..\libs\serialization\src\basic_text_wiprimitive.cpp: Unknown compiler version - please run the configure tests and report the results Error E2061 ..\boost/iterator/iterator_facade.hpp 484: Friends must be functions or classes Error E2139 ..\boost/iterator/iterator_facade.hpp 484: Declaration missing ; Error E2061 ..\boost/iterator/iterator_facade.hpp 485: Friends must be functions or classes Error E2139 ..\boost/iterator/iterator_facade.hpp 485: Declaration missing ; Error E2061 ..\boost/iterator/iterator_facade.hpp 487: Friends must be functions or classes Error E2228 ..\boost/iterator/iterator_facade.hpp 487: Too many error or warning messages *** 6 errors in Compile ***

3 4

Re: [boost] Proposal: Monotonic Containers
by joaquin＠tid.es 09 Jun '09

09 Jun '09

Stewart, Robert escribió: > Andrew Sutton wrote: > On Tuesday, June 09, 2009 9:07 AM > >>> People will expect that a monotonic::foo<..> is like a >>> foo<..>, and they will accept that it requires a storage >>> argument. But they will find it harder to accept that it >>> requires retooling from a type-argument level of the >>> allocator. >>> >> Why incur the overhead of expectation, when foo<T, >> monotonic::allocator<T>> gives you exactly what you want? >> > > There's a problem with using an allocator for the node based containers: they > use the allocator to allocate the elements but not the nodes. That implies > free store (de)allocations for the nodes which is contrary to the intended > purpose of this proposal. > Umm, I beg to differ. If we have a node-based container xxx<T,...,custom_allocator<T> > and node<T> is the internal node type, the container is required to do the allocation using an allocator of type custom_allocator<T>::rebind<node<T> >::other. The container is not allowed to do any kind of allocation using other means than the custom_allocator<X> family of allocators. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

3 4