- Boost - lists.preview.boost.org

Lockfree review
by Phil Endecott 28 Jul '11

28 Jul '11

Dear All, Here is a quick review of the proposed Boost.Lockfree library. Utility ------- For some proposed libraries their utility is beyond doubt because they are already being used in some significant application. In contrast most GSoC projects are developed more in the "hope that they will be useful" without a motivating application. So we have to ask ourselves, does this library meet a need? To me, lock free algorithms and data structures have always seemed appealing at first sight but, apart from some trivial cases like counters, on close inspection they turn out to be too complicated to bother with. (They also lack required features, i.e. waiting for non-empty; I don't think I've ever written a producer-consumer application that didn't need the consumer to block on empty.) The author acknowledges that these structures will typically be slower than conventional implementations using mutexes, but claims that their benefit comes in terms of better worst-case latency. Specifically, if a thread is pre-empted while it holds a lock, another thread will have to wait until the first thread is re-scheduled and can unlock before it can itself lock. But this case is exactly the one that priority inversion is supposed to address: when the second thread tries to lock, the kernel immediately re-schedules the first thread. Yes, because of the kernel involvement this will probably still be slower than the lock-free approach, but perhaps not by much. It seems to me that the author should present benchmarks to demonstrate the actual worst-case latency difference between this code and a lock-based implementation. Design ------ I am happy with the design of the interface, as far as it goes; there are a few things that could be added though: - Debug modes that check the thread requirements. - Other combinations of single/multiple producer/consumer. Implementation -------------- I have looked only at the implementation of the ringbuffer class template. It lacks comments. There are some oddities like this: size_t ret = read_index - write_index - 1; if (write_index >= read_index) ret += max_size; return ret; Here, a negative value may be assigned to the unsigned size_t. I think this is actually safe, but it looks odd. I have not found any bugs. It looks like every call to enqueue() and dequeue() will access both the read and write indexes. Won't this cause poor performance due to cache issues? I would have been tempted to avoid this by keeping local 'available' counts: (PSEUDO_CODE) private: atomic<size_t> read_index_; atomic<size_t> write_index_; size_t read_available_; size_t write_available_; public: bool enqueue(...) { if (write_available_==0) { // In this case, we access read_index_ which is probably cached by the consumer thread, // which is expensive. But this is rare. write_available_ = (read_index_ + max_size - write_index_ - 1) % max_size; } if (!write_available_) return false; // full buffer[write_index_.load(memory_order_relaxed)] = data; write_index_.store((write_index_+1)%max_size); --write_available_; } // dequeue similarly... Documentation ------------- I'm not impressed by the documentation. It lacks rationale and information about the implementation. It has some formatting issues and typos (e.g. lose/loose). Platform requirements (i.e. the Boost.Atomic and/or std::atomic requirements and supported architectures) should be explained. The reference documentation involves more clicks than I would expect to reach the actual content. Some functions are marked as "for debugging only", e.g. the empty() and reset() [which should probably be called clear()] methods. It's not clear to me why these methods must be unsafe; for example, for the ringbuffer, I would expect the consumer thread to safely be able to call empty() and the producer to safely be able to call full() (which is not provided). Rationale for these restrictions (or relaxations of them) would be useful. - Did you try to use the library? With what compiler? Did you have any problems? I considered exercising the library on my ARM systems, but I realised that it requires 64-bit atomics; currently, Boost.Atomic only provides 32-bit atomics on ARM. (This raises the whole issue of the Boost.Atomic dependency. I'm presuming that Tim intends not to bundle a Boost.Atomic snapshot with the library if approved. If that changes, we really need another review.) - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? A few hours reading the docs and code. - Are you knowledgeable about the problem domain? I know a certain amount about atomics and lock-based concurrent programming, but not much about lock-free data structures. - Please always state in your review, whether you think the library should be accepted as a Boost library! I think it should be accepted provided that: * The author demonstrates some measurable benefit over lock-based data structures. * The documentation is improved to describe the implementation and to provide rationale for design decisions. * The library is not added to Boost until Boost.Atomic has been reviewed and approved (or, available std::atomic implementations work with it). Regards, Phil.

2 1

sandbox or boost vault
by Sean Farrow 28 Jul '11

28 Jul '11

Hi: Given that the boost vault has been down for quite a few months, where are all new bosot libraries going for review, in the boost vault-with a different url, or in the boost sandbox in the svn repository? Any help appreciated. Regards Sean.

11 34

[Lockfree review] Meta-comments
by Phil Endecott 28 Jul '11

28 Jul '11

Dear all, I have a few meta comments about the Lockfree review and reviews in general: 1. I know a lot of people read this list via gmane.org. A bug in gmane.org means that when a message is posted to more than one list - for example, a review announcement sent to this list and announce@ - then it appears only in one, and follow-ups are lost. So gmane.org readers will have missed most of this review. Of course a valid answer would be "they should fix it", but considering how few reviews have been written recently I think we should do everything possible to increase discussion. I don't know if review managers get a "how to run your review" document, but assuming that they do, I suggest adding "please post separate announcements to each list, not one message to all". 2. I think every candidate library should have an online copy of its documentation. Again, I think this should be on the review manager's checklist. Yes, you could say that people should be able to download and unzip and/or build their own copy of the docs, but again, we need more reviews and we need to remove as many obstacles (real or imagined) as possible. (In the last review I asked "Where are the online docs?" and it turned out that they were available but I had missed them; oops. If I have once again missed them, oops again.) 3. This has already been discussed to some extent, but why are we reviewing this before Boost.Atomic? I think Boost.Atomic is an urgently-required library. What can we all do to help it get submitted and accepted as soon as possible? Regards, Phil.

8 19

[Lockfree] Tagged pointers
by Phil Endecott 28 Jul '11

28 Jul '11

Tim & others, I understand that your implementation uses tagged pointers to avoid ABA problems. An issue with tagged pointers is that naively they will be 2*sizeof(T*) and systems may not have atomic operations that operate on things of that size. Looking at the code, I see that you also have a "compressed" tagged pointer that is enabled for some architectures. Having found this code, I now vaguely recall some previous mailing list discussions about this. Sometimes it may be possible to claim that users (& reviewers?) of a library do not need to know about details of this sort because the library "just works", and that if they are curious they can read the source. I don't think this is one of those cases. The documentation should explain in outline how this works, and provide rationale for why it is how it is. IIRC, ARM has 64-bit atomics only from architecture version 6k onwards, and my Boost.Atomic ARM code doesn't use them. This could be fixed. Regards, Phil.

2 1

boost threads: pthread mutex priority inheritance
by Jason Newton 28 Jul '11

28 Jul '11

Hi, I just discovered today that pthread mutex's aren't guaranteed to do priority inheritance unless PTHREAD_PRIO_INHERIT is set. I'm doing a bunch of soft-rt work with SCHED_RR so priority inheritance on shared resources is a given necessity. But how should this be worked into boost mutex? There's no way of constructing a boost mutex from an existing object, nor anyway of setting mutex attributes used inside the ctor. There's also no way of manipulating the mutex attributes after pthread_mutex_init. For the moment, the few lines that compose the mutex class are easily copied so that I can do my own business inside the constructor but I expect others will run into this problem down the line and this solution is a bit unsavory. Any better advice?

1 1

[context] MinGW support?
by Nathan Ridge 27 Jul '11

27 Jul '11

Hello, I would like to play around with the (provisionally accepted) Boost.Context library, and I'm wondering: does it support MinGW? First, I should mention, I am using the source from http://ok73.ok.funpic.de/boost.context-0.7.4.zip, but I'm not sure whether this is the latest version - the link to the source at http://www.boost.org/community/review_schedule.html is broken (it points to the Boost Vault, which is down, and I can't find it at https://github.com/boost-vault) I tried to compile it both in 64-bit mode using: toolset=gcc architecture=x86 instruction-set=yorksfield address-model=64 and in 32-bit mode using: toolset=gcc architecture=x86 instruction-set=i686 address-model=32 but in both cases I get errors (pasted below). Am I doing something wrong, or does this library not support MinGW? If the latter, is MinGW support planned at some point? Thanks, Nate 64-bit errors: gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-york sfield\link-static\protected_stack_windows.o libs\context\src\protected_stack_windows.cpp:1:0: sorry, unimplemented: 64-bit mode not compiled in "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -m64 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c - o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfield\link-s tatic\protected_stack_windows.o" "libs\context\src\protected_stack_windows.cpp" ...failed gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instructio n-set-yorksfield\link-static\protected_stack_windows.o... common.mkdir bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfi eld\link-static\asm gcc.compile.asm bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-york sfield\link-static\asm\fcontext_x86_64_ms_pe_masm.o libs\context\src\asm\fcontext_x86_64_ms_pe_masm.asm:1:0: sorry, unimplemented: 64-bit mode not compiled in "g++" -x assembler-with-cpp -O3 -finline-functions -Wno-inline -Wall -m64 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c -o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfield\link-sta tic\asm\fcontext_x86_64_ms_pe_masm.o" "libs\context\src\asm\fcontext_x86_64_ms_pe_masm.asm" ...failed gcc.compile.asm bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instructio n-set-yorksfield\link-static\asm\fcontext_x86_64_ms_pe_masm.o... gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-york sfield\link-static\stack_helper_windows.o libs\context\src\stack_helper_windows.cpp:1:0: sorry, unimplemented: 64-bit mode not compiled in "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -m64 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c - o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfield\link-s tatic\stack_helper_windows.o" "libs\context\src\stack_helper_windows.cpp" ...failed gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instructio n-set-yorksfield\link-static\stack_helper_windows.o... gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-york sfield\link-static\seh.o libs\context\src\seh.cpp:1:0: sorry, unimplemented: 64-bit mode not compiled in "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -m64 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c - o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfield\link-s tatic\seh.o" "libs\context\src\seh.cpp" ...failed gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instructio n-set-yorksfield\link-static\seh.o... ...skipped <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfi eld\link-static>libboost_context-mgw46-1_46_1.a(clean) for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\a ddress-model-64\architecture-x86\instruction-set-yorksfield\link-static>protected_stack_windows.o... ...skipped <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-64\architecture-x86\instruction-set-yorksfi eld\link-static>libboost_context-mgw46-1_46_1.a for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address- model-64\architecture-x86\instruction-set-yorksfield\link-static>protected_stack_windows.o... ...skipped <pstage\lib>libboost_context-mgw46-1_46_1.a for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\a ddress-model-64\architecture-x86\instruction-set-yorksfield\link-static>libboost_context-mgw46-1_46_1.a... ...failed updating 4 targets... ...skipped 3 targets... ...updated 9 targets... 32-bit errors: cc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686 link-static\protected_stack_windows.o ibs\context\src\protected_stack_windows.cpp: In constructor 'boost::contexts::protected_stack::protected_stack(std::siz _t)': ibs\context\src\protected_stack_windows.cpp:60:16: warning: unused variable 'result' [-Wunused-variable] ommon.mkdir bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686\li k-static\asm cc.compile.asm bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686 link-static\asm\fcontext_i386_ms_pe_masm.o ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm: Assembler messages: ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:2: Error: no such instruction: `copyright Oliver Kowalke 2009.' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:3: Error: no such instruction: `distributed under the Boost Software L cense,Version 1.0.' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:4: Error: junk at end of line, first unrecognized character is `(' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:7: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:8: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:9: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:10: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:11: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:12: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:13: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:14: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:15: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:16: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:17: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:18: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:19: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:20: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:21: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:22: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:23: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:24: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:25: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:26: Error: junk at end of line, first unrecognized character is `|' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:27: Error: junk at end of line, first unrecognized character is `-' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:29: Error: unknown pseudo-op: `.386' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:30: Error: unknown pseudo-op: `.xmm' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:31: Error: unknown pseudo-op: `.model' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:32: Error: no such instruction: `_exit PROTO,value:SDWORD' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:33: Error: no such instruction: `boost_fcontext_seh PROTO,except:DWORD frame:DWORD,context:DWORD,dispatch:DWORD' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:34: Error: unknown pseudo-op: `.code' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:36: Error: no such instruction: `boost_fcontext_jump PROC EXPORT' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:37: Error: invalid char '[' beginning operand 2 `[esp+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:37: Error: no such instruction: `load address of the first fcontext_t rg' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:38: Error: invalid char '[' beginning operand 1 `[eax]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:38: Error: no such instruction: `save EDI' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:39: Error: invalid char '[' beginning operand 1 `[eax+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:39: Error: no such instruction: `save ESI' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:40: Error: invalid char '[' beginning operand 1 `[eax+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:40: Error: no such instruction: `save EBX' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:41: Error: invalid char '[' beginning operand 1 `[eax+0ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:41: Error: no such instruction: `save EBP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:43: Error: no such instruction: `assume fs:nothing' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:44: Error: too many memory references for `mov' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:44: Error: no such instruction: `load NT_TIB' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:45: Error: no such instruction: `assume fs:error' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:46: Error: invalid char '[' beginning operand 2 `[edx]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:46: Error: no such instruction: `load current SEH exception list' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:47: Error: invalid char '[' beginning operand 1 `[eax+02ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:47: Error: no such instruction: `save current exception list' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:48: Error: invalid char '[' beginning operand 2 `[edx+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:48: Error: no such instruction: `load current stack base' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:49: Error: invalid char '[' beginning operand 1 `[eax+020h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:49: Error: no such instruction: `save current stack base' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:50: Error: invalid char '[' beginning operand 2 `[edx+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:50: Error: no such instruction: `load current stack limit' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:51: Error: invalid char '[' beginning operand 1 `[eax+024h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:51: Error: no such instruction: `save current stack limit' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:52: Error: invalid char '[' beginning operand 2 `[edx+010h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:52: Error: no such instruction: `load fiber local storage' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:53: Error: invalid char '[' beginning operand 1 `[eax+030h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:53: Error: no such instruction: `save fiber local storage' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:55: Error: invalid char '[' beginning operand 1 `[eax+018h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:55: Error: no such instruction: `save SSE2 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:56: Error: invalid char '[' beginning operand 1 `[eax+01ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:56: Error: no such instruction: `save x87 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:58: Error: invalid char '[' beginning operand 2 `[esp+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:58: Error: no such instruction: `exclude the return address' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:59: Error: invalid char '[' beginning operand 1 `[eax+010h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:59: Error: no such instruction: `save as stack pointer' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:60: Error: invalid char '[' beginning operand 2 `[esp]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:60: Error: no such instruction: `load return address' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:61: Error: invalid char '[' beginning operand 1 `[eax+014h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:61: Error: no such instruction: `save return address' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:64: Error: invalid char '[' beginning operand 2 `[esp+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:64: Error: no such instruction: `load address of the second fcontext_t arg' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:65: Error: invalid char '[' beginning operand 2 `[eax]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:65: Error: no such instruction: `restore EDI' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:66: Error: invalid char '[' beginning operand 2 `[eax+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:66: Error: no such instruction: `restore ESI' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:67: Error: invalid char '[' beginning operand 2 `[eax+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:67: Error: no such instruction: `restore EBX' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:68: Error: invalid char '[' beginning operand 2 `[eax+0ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:68: Error: no such instruction: `restore EBP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:70: Error: no such instruction: `assume fs:nothing' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:71: Error: too many memory references for `mov' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:71: Error: no such instruction: `load NT_TIB' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:72: Error: no such instruction: `assume fs:error' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:73: Error: invalid char '[' beginning operand 2 `[eax+02ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:73: Error: no such instruction: `load SEH exception list' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:74: Error: invalid char '[' beginning operand 1 `[edx]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:74: Error: no such instruction: `restore next SEH item' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:75: Error: invalid char '[' beginning operand 2 `[eax+020h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:75: Error: no such instruction: `load stack base' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:76: Error: invalid char '[' beginning operand 1 `[edx+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:76: Error: no such instruction: `restore stack base' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:77: Error: invalid char '[' beginning operand 2 `[eax+024h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:77: Error: no such instruction: `load stack limit' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:78: Error: invalid char '[' beginning operand 1 `[edx+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:78: Error: no such instruction: `restore stack limit' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:79: Error: invalid char '[' beginning operand 2 `[eax+030h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:79: Error: no such instruction: `load fiber local storage' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:80: Error: invalid char '[' beginning operand 1 `[edx+010h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:80: Error: no such instruction: `restore fiber local storage' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:82: Error: invalid char '[' beginning operand 1 `[eax+018h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:82: Error: no such instruction: `restore SSE2 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:83: Error: invalid char '[' beginning operand 1 `[eax+01ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:83: Error: no such instruction: `restore x87 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:85: Error: invalid char '[' beginning operand 2 `[eax+010h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:85: Error: no such instruction: `restore ESP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:86: Error: invalid char '[' beginning operand 2 `[eax+014h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:86: Error: no such instruction: `fetch the address to return to' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:88: Error: too many memory references for `xor' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:88: Error: no such instruction: `set EAX to zero' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:89: Error: no such instruction: `indirect jump to context' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:90: Error: no such instruction: `boost_fcontext_jump ENDP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:92: Error: no such instruction: `boost_fcontext_make PROC EXPORT' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:93: Error: invalid char '[' beginning operand 2 `[esp+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:93: Error: no such instruction: `load address of the fcontext_t arg0' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:94: Error: invalid char '[' beginning operand 1 `[eax]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:94: Error: no such instruction: `save the address of current context' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:95: Error: invalid char '[' beginning operand 2 `[esp+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:95: Error: no such instruction: `load the address of the function supp sed to run' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:96: Error: invalid char '[' beginning operand 1 `[eax+014h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:96: Error: no such instruction: `save the address of the function supp sed to run' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:97: Error: invalid char '[' beginning operand 2 `[eax+020h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:97: Error: no such instruction: `load the stack base' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:98: Error: invalid char '[' beginning operand 2 `[edx-014h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:98: Error: no such instruction: `reserve space for last frame on stack (ESP+4)% 16==0' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:99: Error: invalid char '[' beginning operand 1 `[eax+010h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:99: Error: no such instruction: `save the address' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:101: Error: too many memory references for `mov' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:101: Error: no such instruction: `set ECX to exception-handler' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:102: Error: invalid char '[' beginning operand 1 `[edx+0ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:102: Error: no such instruction: `save ECX as SEH handler' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:103: Error: too many memory references for `mov' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:103: Error: no such instruction: `set ECX to -1' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:104: Error: invalid char '[' beginning operand 1 `[edx+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:104: Error: no such instruction: `save ECX as next SEH item' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:105: Error: invalid char '[' beginning operand 2 `[edx+08h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:105: Error: no such instruction: `load address of next SEH item' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:106: Error: invalid char '[' beginning operand 1 `[eax+02ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:106: Error: no such instruction: `save next SEH' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:108: Error: invalid char '[' beginning operand 2 `[eax+028h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:108: Error: no such instruction: `load the address of the next context ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:109: Error: invalid char '[' beginning operand 1 `[eax+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:109: Error: no such instruction: `save the address of the next context ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:110: Error: invalid char '[' beginning operand 2 `[esp+0ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:110: Error: no such instruction: `load the address of the void pointer arg2' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:111: Error: invalid char '[' beginning operand 1 `[edx+04h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:111: Error: no such instruction: `save the address of the void pointer onto the context stack' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:112: Error: invalid char '[' beginning operand 1 `[eax+018h]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:112: Error: no such instruction: `save SSE2 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:113: Error: invalid char '[' beginning operand 1 `[eax+01ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:113: Error: no such instruction: `save x87 control word' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:114: Error: too many memory references for `mov' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:114: Error: no such instruction: `load helper code executed after fn() eturns' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:115: Error: invalid char '[' beginning operand 1 `[edx]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:115: Error: no such instruction: `save helper code executed adter fn() eturns' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:116: Error: too many memory references for `xor' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:116: Error: no such instruction: `set EAX to zero' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:118: Error: no such instruction: `boost_fcontext_make ENDP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:120: Error: no such instruction: `boost_fcontext_link PROC' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:121: Error: invalid char '[' beginning operand 2 `[esp-0ch]' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:121: Error: no such instruction: `adjust the stack to proper boundarie ' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:122: Error: too many memory references for `test' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:122: Error: junk `a next context was given' after expression ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:122: Error: number of operands mismatch for `test' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:123: Error: no such instruction: `jump to finish' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:125: Error: junk `address of the next context on the stack' after expr ssion ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:126: Error: junk `address of the current context on the stack' after e pression ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:127: Error: no such instruction: `install next context' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:130: Error: too many memory references for `xor' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:130: Error: no such instruction: `set EAX to zero' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:131: Error: no such instruction: `exit code is zero' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:132: Error: no such instruction: `exit application' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:134: Error: no such instruction: `boost_fcontext_link ENDP' ibs\context\src\asm\fcontext_i386_ms_pe_masm.asm:135: Error: no such instruction: `end' "g++" -x assembler-with-cpp -O3 -finline-functions -Wno-inline -Wall -march=i686 -m32 -DBOOST_ALL_NO_LIB=1 -DNDEBUG I"." -c -o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686\li k-static\asm\fcontext_i386_ms_pe_masm.o" "libs\context\src\asm\fcontext_i386_ms_pe_masm.asm" ..failed gcc.compile.asm bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instructio -set-i686\link-static\asm\fcontext_i386_ms_pe_masm.o... cc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686 link-static\stack_helper_windows.o cc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686 link-static\seh.o ibs\context\src\seh.cpp: In function 'char* exception_description(const _EXCEPTION_RECORD*, char*, size_t)': ibs\context\src\seh.cpp:21:40: error: '_TRUNCATE' was not declared in this scope ibs\context\src\seh.cpp:21:100: error: '_snprintf_s' was not declared in this scope ibs\context\src\seh.cpp:24:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:25:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:26:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:27:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:28:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:29:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:30:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:31:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:32:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:33:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:34:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:35:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:36:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:37:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:38:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:39:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:40:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:41:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:42:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:43:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:44:53: warning: deprecated conversion from string constant to 'char*' [-Wwrite-strings] ibs\context\src\seh.cpp:47:36: error: '_TRUNCATE' was not declared in this scope ibs\context\src\seh.cpp:47:71: error: '_snprintf_s' was not declared in this scope ibs\context\src\seh.cpp: At global scope: ibs\context\src\seh.cpp:52:1: error: 'EXCEPTION_DISPOSITION' does not name a type ibs\context\src\seh.cpp:9:15: warning: 'char* exception_description(const _EXCEPTION_RECORD*, char*, size_t)' defined b t not used [-Wunused-function] "g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -march=i686 -m32 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c -o "bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686\ ink-static\seh.o" "libs\context\src\seh.cpp" ..failed gcc.compile.c++ bin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instructio -set-i686\link-static\seh.o... ..skipped <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686\li k-static>libboost_context-mgw46-1_46_1.a(clean) for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address model-32\architecture-x86\instruction-set-i686\link-static>asm\fcontext_i386_ms_pe_masm.o... ..skipped <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model-32\architecture-x86\instruction-set-i686\li k-static>libboost_context-mgw46-1_46_1.a for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\address-model- 2\architecture-x86\instruction-set-i686\link-static>asm\fcontext_i386_ms_pe_masm.o... ..skipped <pstage\lib>libboost_context-mgw46-1_46_1.a for lack of <pbin.v2\libs\context\build\gcc-mingw-4.6.0\release\a dress-model-32\architecture-x86\instruction-set-i686\link-static>libboost_context-mgw46-1_46_1.a... ..failed updating 2 targets... ..skipped 3 targets... ..updated 7 targets...

3 12

BOOST_CLASS_EXPORT in C++
by Eric Maes 27 Jul '11

27 Jul '11

Hello all, I am new to the boost community and libraries and need some help. I am taking over a small project which uses boost serialization. We are having trouble using the BOOST_CLASS_EXPORT macro with version 1.39. We are in the process of upgrading to v1.47, but have not yet complete the transition on our other supporting DLLs. My question is this: When we include the macro in our CPP file, it builds ok and appears to be ready to rock, however when we attempt to serialze the object, an exception is thrown. after stepping deeply, we see that the derived class load_construct_data and save_construct_data methods are not getting called. We have followed the documentation and it appears we have it implemented properly. Shown here are code snippets of what we are doing... When the macro use was in the HPP files, the serialization worked, but then an exception was thrown when the application ends. Either way, there is an issue. From what I can tell in boost docs, the macro use needs to be in the CPP and not in the HPP If anyone can give assistance that would be great. I am so new to Boost that I'm not sure I am using the right nomanclature to describe what is occuring. Any help that I can get will be much appreciated. Eric emaes(a)miengsrv.com // ************ derived class (Calculator) load/save definition ********* // template inline void save_construct_data (Archive& ar, const LpsTktMDO::Calculator* t, const unsigned int file_version) { // save data required to construct instance Lps::CharString name; Lps::CharString expression; name=t->getName (); expression=t->getExpression (); ar << boost::serialization::make_nvp ("name", name); ar << boost::serialization::make_nvp ("expression", expression); } //! //! Serialization method for loading a Calculator object //! //! \return None //! template inline void load_construct_data (Archive& ar, LpsTktMDO::Calculator* t, const unsigned int file_version) { // retrieve data from archive required to construct new instance Lps::CharString name; Lps::CharString expression; ar >> boost::serialization::make_nvp ("name", name); ar >> boost::serialization::make_nvp ("expression", expression); // invoke inplace constructor to initialize instance of my_class ::new(t)LpsTktMDO::Calculator (name, expression); } // ************ base class (Function) serialize definition ********* // template void serialize (Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP (theNameM); ar & BOOST_SERIALIZATION_NVP (thePmmIdM); ar & BOOST_SERIALIZATION_NVP (theDataMgrM); }; // ************ call to serialize the map of calculator/function objects (which is a member of a PreDef object) ********* // template void serialize (Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP (theDesignVariablesM); // <<-------- this works ar & BOOST_SERIALIZATION_NVP (theVariablesM); // <<-------- this works ar & BOOST_SERIALIZATION_NVP (theFunctionsM); // <<-------- this call issues the exception! ar & BOOST_SERIALIZATION_NVP (theDataMgrM); // <<-------- this works }; // ************ call to serialize the PreDef object ********* // BOOST_CHECK_NO_THROW(theOa << BOOST_SERIALIZATION_NVP(thePreDefM)); -- View this message in context: http://boost.2283326.n4.nabble.com/BOOST-CLASS-EXPORT-in-C-tp3696005p369600… Sent from the Boost - Dev mailing list archive at Nabble.com.

2 2

[lockfree] review
by Artyom Beilis 27 Jul '11

27 Jul '11

Hello, Please consider it as a 1/2 or 1/3 review as I hadn't much time to review this library: I'd start answering the questions in little bit different order: ------------------------------------------------------ - What is your evaluation of the design? Basically it looks quite simple and straight forward. As the data structures are based (according to the author) on the current papers and studies; it uses quite standard features and seems fine. ------------------------------------------------------ - What is your evaluation of the implementation? It seems to be clear, however I miss following: As it is virtually impossible to test concurrent data structures I expect that every critical part in the code that describes concurrent operation would have a reference to specific article or book and include full description and even a small proof that it is right taken from the resources. I think it is very critical for maintenance of the library and in ensuring the fact that it can be maintainable. Finally it looks like a solid code written with good knowledge of what is going on. ------------------------------------------------------ - What is your evaluation of the documentation? I think it is too brief and misses important parts. For example: html/boost/lockfree/ringbuffer.html Should include in the reference general requirements like it should be single writer single reader and not mentioned in some other places. In the first glance I assumed that it was multiple-readers-writers ring buffer. ------------------------------------------------------ - Did you try to use the library? With what compiler? Did you have any problems? I've tested it with gcc-4.3 on cygwin using CMake scripts,no particular problems. ------------------------------------------------------ - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I've spend 2-3 hours reading the code, comparing my own "ringbuffer" with a provided ring buffer and fifo. ------------------------------------------------------ - What is your evaluation of the potential usefulness of the library? ok... This is probably the biggest problem with this library. I understand that Linked List, Stack, Ring are actually the implementable data structures (most of others are much harder to do if possible at all, even thou there are few that can be implemented: hash tables and some others) I understand that this is "lock-free" library not generic-concurrent-data-structures library but this is why I think that it may not be as useful as it seems. I'd like to see more: 1. Queues with OS support like waiting - non-waiting with lock-free access when the queue not-empty and not-full 2. Queues with different optimization for different cases. 3. There are several ways to optimize stacks to reduce contention. 4. Some transactional memory blocks that allow to implement optimistic strategies 5. Combination of locking and lock-free parts. 6. Hash tables And more. ------------------------------------------------------ - Are you knowledgeable about the problem domain? Yes: - I had implemented several concurrent data structures for professional needs. - I had took Prof. Nir Shavit's course in concurrent programming (The author of the The Art of Multiprocessor Programming) ------------------------------------------------------ Should the library be accepted into Boost? I have doubts, mostly because my considerations about the usability of the library in the real applications. I'd expect the author to provide more data structures more available policies, for example there are several ways to optimize stack object see "The Art of Multiprocessor Programming" chapter 11 section 4. Also I'd like to see some more helpful objects in the interface: Tagged ptr is something that would be seen for many data structures I don't think should it be in details? I really hesitate between "Conditional YES" and "YES" - Finally it seems to be doing what it does - It is not concurrent data structures library but lock-free one. So I vote "Yes" but I strongly recommend and expect from the author: - Make some building blocks available in detains to library interface allowing users to implement other data-structures easier. - Improve documentation about what and how it should be done - Provide more data structures in future and provide data structures with different policies Artyom Beilis -------------- CppCMS - C++ Web Framework: http://cppcms.sf.net/ CppDB - C++ SQL Connectivity: http://cppcms.sf.net/sql/cppdb/

3 2

Boost.Locale 'Status'?
by Joshua Boyce 27 Jul '11

27 Jul '11

I was wondering what the 'status' of the Boost.Locale library is now that it has been accepted into Boost. I can't see it on the SVN, what are the current plans for its integration into mainline Boost? Thanks.

4 12

[chrono] gcc visibility support patch
by Jürgen Hunold 27 Jul '11

27 Jul '11

Hi, I've found some missing patches for https://svn.boost.org/trac/boost/ticket/2114 This one enables visibility support for Boost.Chrono For more information, please see http://gcc.gnu.org/wiki/Visibility Okay to commit? Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold(a)ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !

2 2

[Opaque] Request for interest in Opaque typedefs library emulation
by vicente.botet 27 Jul '11

27 Jul '11

Hi, I have started to play with a class to define opaque typedefs on funcdamental types. It is based on the C++ proposal " Progress toward Opaque Typedefs for C++0X" by Walter E. Brown http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1891.pdf and http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1706.pdf Please take a look on the preceding links to understand the motivation. Boost.Sericalization provide a BOOST_STRONG_TYPEDEF already, but in my opinon the implementation is not complete,as only the order operators are defined (But maybe I'm wrong). Evidently it can not provide everything we can have with a language approach, but IMO this could yet be useful. The library include two class templates (CRTP) to define opaque types: * public_opaque_type<OT,UT> : implicit conversion from OT to UT * private_opaque_type<OT,UT> : no conversion from OT to UT An opaque type (OT) is a new type, different from the underlying type (UT) and other OT. It defines explicit constructor from the UT, and defines all the operators the UT support but this time applied to the OT. To define a new OT with this mixin classes struct serial_number: public_opaque_type<serial_number, unsigned> { BOOST_OPAQUE_PUPLIC_FORWARD_CONSTRUCTORS(serial_number,unsigned); }; struct game_score: private_opaque_type<game_score, unsigned> { BOOST_OPAQUE_PRIVATE_FORWARD_CONSTRUCTORS(game_score,unsigned); }; In addition, defines macros that make this simpler BOOST_OPAQUE_PUBLIC_TYPEDEF(unsigned, serial_number); BOOST_OPAQUE_PRIVATE_TYPEDEF(unsigned, game_score); Next follows the definition of the public_opaque_type template <typename Final, typename T> class public_opaque_type : boost::totally_ordered< Final , boost::integer_arithmetic< Final , boost::bitwise< Final , boost::unit_steppable< Final , boost::totally_ordered< T, Final // public specific conversions , boost::integer_arithmetic< T, Final // public specific conversions , boost::bitwise< T, Final // public specific conversions > > > > > > > { protected: T val_; typedef public_opaque_type opaque_type; public: typedef T underlying_type; explicit public_opaque_type(const T v) : val_(v) {}; public_opaque_type() {}; public_opaque_type(const opaque_type & rhs) : val_(rhs.val_) {} public_opaque_type & operator=(const opaque_type & rhs) { val_ = rhs.val_; return *this; } public_opaque_type & operator=(const T rhs) { val_ = rhs; return *this; } T const& underlying() const { return val_; } T& underlying() { return val_; } operator const T & () const { return val_; } operator T & () { return val_; } #if 0 bool operator==(const opaque_type & rhs) const { return val_ == rhs.val_; } #endif bool operator<(const Final & rhs) const { return val_ < rhs.val_; } Final& operator+=(const Final & rhs) { val_ += rhs.val_; return static_cast<Final&>(*this); } Final& operator-=(const Final & rhs) { val_ -= rhs.val_; return static_cast<Final&>(*this); } ... }; I have started to make some tests and it works well for the moment. Limitations: * As we can not specialize the behaviour of static_cast/dynamic_cast the library sould provide two specific casts opaque_static_cast and opaque_dynamic_static_cast (Not yet implemented). * As operator.() can not be overloaded, it works transparently only for underlying types that don't defines member functions (C++ fundamental types, as int, char, ...). For the others an explicit use of the underlying() member function is needed :( * Others I have not catched yet :( Do you think it is worth continuing? Please, let me know if you see a error on the design. _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/

4 8

[1.48.0] Proposed Release Schedule
by Beman Dawes 27 Jul '11

27 Jul '11

Give we are already in the middle of July, I feel strongly 1.48.0 should target the first Monday in November, thus putting us back on our usual quarterly cycle. In our earlier schedule discussion, there was some interest in a three times a year release schedule, but the arguments were not convincing to me, so I really think we should keep to our quarterly schedule. The current formula for intermediate dates goes like this: * One week after prior release ships: branches/release opens for merging all stable changes, including bug fixes, and major upgrades to existing libraries. Breaking changes should be coordinated with libraries affected. New libraries may be added with permission of a release manager. * Seven weeks before release: branches/release closed for new libraries and breaking changes to existing libraries. Still open for bug fixes and other routine changes to all libraries. * Six weeks before release: QA checks on snapshot doc builds, inspect status, getting started guide, and install. * Four weeks before release: branches/release closed for major code changes. Still open for serious problem fixes and docs changes. * Three weeks before release: branches/release closed for all library changes except when specific permission given by release manager(s). * Two weeks before release: Beta target date. Further betas and/or release candidates as feedback dictates. I propose that the Beta be scheduled for *Four* weeks before the release target date. All the other intermediate dates would thus be two weeks earlier. That's more or less what we did for 1.47.0 and it allowed enough time to fix a bunch of bugs in the beta that were bothering users. --Beman

10 13

Lockfree impl comment - does not seem to be compatible with Boost::Interprocess (shared memory) [SEC=UNCLASSIFIED]
by Levy, Gabriel (Contractor) 27 Jul '11

27 Jul '11

UNCLASSIFIED Hi, Since Lockfree library (by Tim Blechmann) is in the review stage I have a comment to make. I only looked at the code just at a glance to check if it would be able to be used with Boost::Interprocess library, especially to be placed in the shared memory and to me it definitely doesn't seem to deduce the pointer types from the Allocation supplied as a template argument. For any container implementation to be useable with Interprocess shared memory the requirement is that any internal pointer types are not explicitly stated - instead they aggregate the pointer types from the Allocator's internal pointer typedefs. By this the container will use "smart" or plain pointers, whichever the Allocators specify/define. Good place to look at for how this should be done is <boost/interprocess/containers/map.hpp> and other containers' implementation in that location. Also the Interprocess documentation explains why, for instance, the std::map or std::vector containers are incorrectly (not generically enough) implemented and suggests a possibility of potential future corrections in STL implementations (that part is only speculative, of course). While in its current form it is already very useful in intra-process applications, I would consider the incompatiblity with Interprocess' shared memory a hugely missed opportunity. My recommendation would be to first to address this issue before accepting it for inclusion. Thank you all so much for the Boost - the best thing that happened to C++. Kindest regards Gabe Gabe Levy Senior Software Engineer, Contractor, Estrings Pty Ltd IMPORTANT: This email remains the property of the Department of Defence and is subject to the jurisdiction of section 70 of the Crimes Act 1914. If you have received this email in error, you are requested to contact the sender and delete the email.

2 1

shared_ptr<>: why no operator[] ?
by dgwsoft 27 Jul '11

27 Jul '11

boost::shared_ptr<int> p(new int(42)); To access elements of the array I can do: p.get()[10] = 7; But wouldn't it be nicer to do: p[10] = 7; // ? So: why is no T& operator[](int) defined for shared_ptr ? I'm just interested in the design reasons for this. cheers Gareth -- View this message in context: http://boost.2283326.n4.nabble.com/shared-ptr-why-no-operator-tp3697208p369… Sent from the Boost - Dev mailing list archive at Nabble.com.

4 5

[Review] Lockfree review: two more days left
by Hartmut Kaiser 26 Jul '11

26 Jul '11

Hi all, We have two more days left for the review of Tim Blechmann's Lockfree library. If you're interested in this library, please consider contributing a review! Please see the original (slightly corrected) announcement below: ----------------------------------------------------- About the library: Boost.Lockfree provides implementations of lock-free data structures. Lock-free data structures can be accessed by multiple threads without the necessity of blocking synchronization primitives such as guards. Lock-free data structures can be used in real-time systems, where blocking algorithms may lead to high worst-case execution times, to avoid priority inversion, or to increase the scalability for multi-processor machines. The following data structures are provided: - boost::lockfree::fifo, a lock-free fifo queue - boost::lockfree::stack, a lock-free stack - boost::lockfree::ringbuffer, a wait-free single-producer/single-consumer ringbuffer. The library is accessible from here: http://tim.klingt.org/boost_lockfree.tar.gz. Boost.Lockfree depends on C++0x atomics. They are not well supported with current compilers yet. For this reason Boost.Lockfree depends on the *unreviewed* Boost.Atomic library, that emulates C++0x atomics for C++98. This review is about Boost.Lockfree and not about Boost.Atomic, although it is included in the tarball and the git repository. If Boost.Lockfree will be accepted, it won't be merged into trunk before Boost.Atomic is accepted. --------------------------------------------------- Please always state in your review, whether you think the library should be accepted as a Boost library! Additionally please consider giving feedback on the following general topics: - What is your evaluation of the design? - What is your evaluation of the implementation? - What is your evaluation of the documentation? - What is your evaluation of the potential usefulness of the library? - Did you try to use the library? With what compiler? Did you have any problems? - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - Are you knowledgeable about the problem domain? Regards Hartmut Review Manager

3 2

[range] Should ranges really propagate constess to the data ?
by Samuel Debionne 26 Jul '11

26 Jul '11

Hello everyone, Here is an example code that motivates my question : typedef std::vector<int> vector_t; boost::sub_range<vector_t> three_first(vector_t& _v) {return boost::make_iterator_range(_v.begin(), _v.begin() + 3); } int main(int argc, char* argv[]) { vector_t v; boost::fill(three_first(v), 1); // Fail : you cannot assign to a variable that is const return 0; } Taken from Boost.GIL documentation, the definition of image views is : An image view is a generalization of STL's range concept to multiple dimensions. Similar to ranges (and iterators), image views are shallow, don't own the underlying data and don't propagate their constness over the data. For example, a constant image view cannot be resized, but may allow modifying the pixels. For pixel-immutable operations, use constant-value image view (also called non-mutable image view). I expected the range concept to follow these lines regarding constness, but I'm not very sure anymore... for instance sub_range have : iterator begin() { return base::begin(); } *const_iterator* begin() const { return base::begin(); } and boost::begin() is : inline range_iterator<T>::type begin( T& r ) {...} inline range_iterator<*const* T>::type begin( const T& r ) {...} which means that I'm probably wrong. Am I missing something ? Regards, Samuel

7 15

[regex] How robust are the <boost/regex/pending/unicode_iterator.hpp> adapters?
by Beman Dawes 26 Jul '11

26 Jul '11

Boost.Filesystem needs the UTF-32 to UTF-16 and UTF-16 to UTF-32 adapters to implement char16_t and char32_t support. Do they have any known bugs or other outstanding problems? TIA, --Beman

7 20

Re: [boost] [range] Should ranges really propagate constess to the data ?
by Brian Wood 25 Jul '11

25 Jul '11

Date: Mon, 25 Jul 2011 10:01:09 +0200 From: Thorsten Ottosen <thorsten.ottosen(a)dezide.com> To: boost(a)lists.boost.org Subject: Re: [boost] [range] Should ranges really propagate constess to the data ? Message-ID: <4E2D22C5.7050701(a)dezide.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Thorsten Ottosen: >Den 24-07-2011 16:04, Mathias Gaunard skrev: >> >> I thought that was the point of sub_range: to be the same as >> iterator_range, except the template parameter it takes is a range rather >> than an iterator, simplifying things when only ranges and not iterators >> are involved. >> > > + propagation of constness. > > Consider using sub_range<T> as the type of a member variable. > > It would be very annoying in some cases that we cannot return a view > that is const-correct and which requires us to construct a temporary > view instead of just returning a reference to the existing view. > The divergent opinions on sub_range in this thread convinced me to add marshalling support for iterator_range to the code generator I'm working on and I've now done so. Previously I had only added support for sub_range. Regards, Brian Wood Ebenezer Enterprises http://webEbenezer.net

1 0

[quickbook] Revision 73337 doesn't compile with msvc-10.0
by Robert Kawulak 25 Jul '11

25 Jul '11

Here's the relevant part of the output: S:\Program Files\Microsoft Visual Studio 10.0\VC\INCLUDE\deque(795) : error C202 7: use of undefined type 'quickbook::id_generator::placeholder' s:\program files\boost_trunk\tools\quickbook\src\id_generator.hpp(36) : see declaration of 'quickbook::id_generator::placeholder' s:\program files\boost_trunk\tools\quickbook\src\id_generator.hpp(40) : see reference to class template instantiation 'std::deque<_Ty>' being compiled with [ _Ty=quickbook::id_generator::placeholder ] Best regards, Robert

1 0

[quickbook] Template issues
by Robert Kawulak 25 Jul '11

25 Jul '11

Hi, I have some problems using QuickBook templates. This code: [article test [quickbook 1.5]] [template link_example[name] [@[name] [name]] ] [link_example some_link] Generates: <ulink url="[name">[name</ulink> some_link] Which is obviously not what's expected. The second issue is with the code: [article test [quickbook 1.5]] [template import_example[name] [import ../temp/[name].cpp] [[name]] ] [import_example example] This generates the following error: Generating Output File: test.xml ../temp/[name: error: Could not open input file. Error: Error count: 1. Are these the fault of QuickBook or am I doing something wrong? (I'm using trunk revision 73337.) Thanks and best regards, Robert

1 0

[ggl] Congratulations
by Simonson, Lucanus J 25 Jul '11

25 Jul '11

Barend, I wanted to congratulate you on the release of your library as a part of boost. I know how hard and how long you've worked for this and your efforts are commendable. The community you have built around the library is healthy, vibrant and growing. I wish you the best continuing success. Regards, Luke

2 1

[lexical_cast] broken for std::vector<char> since last updates
by Jürgen Hunold 25 Jul '11

25 Jul '11

Hi, the last patches to lexical_cast broke some of my code which uses lexical_cast to parse values from std::vector<char> Something as simple as std::vector<char> buffer; return boost::lexical_cast<int>(&buffer[0]); fails with BOOST_ROOT/boost/lexical_cast.hpp:1091:43: error: no matching constructor for initialization of 'std::basic_ostream<char>' std::basic_ostream<CharT> stream(&stringbuffer); ^ ~~~~~~~~~~~~~ this is the start of the clang diagnostic. Both gcc (Ubuntu 4.6.0-3~ppa1) 4.6.1 20110409 (prerelease) and clang version 3.0 (trunk 135897) output contained in attached testcase. Btw. are there any lexical tests in Boost? Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold(a)ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !

3 4

[locale] UTF Iteration and conversion
by Artyom Beilis 25 Jul '11

25 Jul '11

Hello All, Following this discussion: [regex] How robust are the <boost/regex/pending/unicode_iterator.hpp> adapters? http://thread.gmane.org/gmane.comp.lib.boost.devel/221643 Boost.Locale will ship **header-only** utf_to_utf conversion functions and provide operations that would allow implementing iteration over UTF-XX sequences. This was one of the review requests and it will be integrated in few days into trunk. Functions: namespace boost { namespace locale { namespace conv { template<typename CharOut,typename CharIn> std::basic_string<CharOut> conv::utf_to_utf(CharIn const *begin,CharIn const *end,method_type how=skip); template<typename CharOut,typename CharIn> std::basic_string<CharOut> conv::utf_to_utf(CharIn const *str,method_type how=skip); template<typename CharOut,typename CharIn> std::basic_string<CharOut> conv::utf_to_utf(std::basic_string<CharIn> const &str,method_type how=skip); }}} // boost::locale::conv Also it would provide basic utf encoding/decoding operations that would allow to create iterators over generic streams namespace boost { namespace locale { namespace utf { typedef uint32_t code_point; static const code_point illegal = ... ; static const code_point incomplete = ... ; // Get one code point from UTF stream template<typename InputIterator> code_point decode(InputIterator &p,InputIterator e); // Write one code point to UTF stream template<typename OutputIterator> Iterator encode(code_point u,OutputIterator e); // Get size of code point in basic unit template<typename CharType> int units_in_code_point(code_point u); }}} // boost::locale::utf It should be ready very soon... Artyom Beilis -------------- CppCMS - C++ Web Framework: http://cppcms.sf.net/ CppDB - C++ SQL Connectivity: http://cppcms.sf.net/sql/cppdb/

1 0

[asio] gcc visibility support patch
by Jürgen Hunold 25 Jul '11

25 Jul '11

Hi, I've found some missing patches for https://svn.boost.org/trac/boost/ticket/2114 This one enables visibility support for Boost.Asio For more information, please see http://gcc.gnu.org/wiki/Visibility Okay to commit? Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! Ingenieurgesellschaft für * voice: ++49 511 262926 57 ! Verkehrs- und Eisenbahnwesen mbH * fax : ++49 511 262926 99 ! Lister Straße 15 * juergen.hunold(a)ivembh.de ! www.ivembh.de * * Geschäftsführer: ! Sitz des Unternehmens: Hannover * Prof. Dr.-Ing. Thomas Siefer ! Amtsgericht Hannover, HRB 56965 * PD Dr.-Ing. Alfons Radtke !

3 5

[lockfree review] rfc: naming and interface
by Tim Blechmann 25 Jul '11

25 Jul '11

hi all, i am starting a new thread, because i want to start a new discussion on one aspect of boost.lockfree: the naming of the data structure and the interface naming: the current names of the data structures are fifo, stack and ringbuffer. during the pre-review some people suggested to use different names, lifo instead of stack or queue instead of fifo. in general i think it is a good idea to use a consistent naming (probably fifo should be renamed to queue), but how does the ringbuffer come in? it is a queue as well, but with different characteristics (spsc and wait-free). data structure configuration: stack and fifo currently use 3 template arguments: T for the managed type, freelist_t as a tag to select the underlying freelist and Alloc as the allocator which is used for the internal nodes: template<typename T, typename freelist_t = caching_freelist_t, typename Alloc = std::allocator<T> > class fifo/stack; however i have been thinking about replacing it with a boost.parameter interface, something like: template<typename T, ... Options> class fifo/stack; with the options: boost::lockfree::freelist_can_allocate<true/false> to disable memory allocation during enqueue/push boost::lockfree::allocator<> to specify the allocator ringbuffer size: the ringbuffer currently has the signature: template<typename T, size_t max_size> class ringbuffer; if max_size is 0, the size of the ringbuffer can be configured at run-time with a constructor. this is not really the cleanest interface, but i would like to provide an interface, which supports both compile-time and run-time setting of the size. i would apprechiate a discussion about these points during the review. thanks, tim

7 10