June 2020 - Boost - lists.preview.boost.org

[epochs] Boost conditional dependency map
by Joaquin M López Muñoz 05 Jun '20

05 Jun '20

As part of the epoch proposal, I've written a small toolchain to generate a so-called Boost conditional dependency map: https://github.com/joaquintides/boostcdmap/ Such a map gives, for each Boost module and each epoch Boost03, ... , Boost20, the dependencies of the module under the given epoch (which may vary across epochs). The actual map for Boost 1.73 is provided at: https://github.com/joaquintides/boostcdmap/blob/master/boostcdmap.1.73.0.js… The conditional dependency map is a key asset for the automation of epoch assignment; I'll elaborate on this point in the upcoming days. As always, comments welcome. Best regards, Joaquín M López Muñoz

1 0

[epochs] Internal relevance in post-C++03 Boost of Boost03 components vs. standard equivalents
by Joaquin M López Muñoz 05 Jun '20

05 Jun '20

For those of you who are following the discussion on epochs, I've just prepared a report on the relevance of "legacy" components intra-Boost for these Boost libs that require C++11 (or higher) to work: https://github.com/joaquintides/boost_epoch/blob/master/boost_vs_std_intern… Quoting from the report: "It can be seen that (direct) usage of Boost03 components is residual (and in some cases due to backwards compatibility rather than for their functionality). We can draw then the conclusion that newer Boost libraries are already dismissing legacy dependencies when the standard environment offers alternatives." Best, Joaquín M López Muñoz

2 3

[LEAF] A sort of review
by Andrzej Krzemienski 04 Jun '20

04 Jun '20

Hi Everyone, According to the schedule the review period of LEAF should have ended, but since I haven't heard a clear announcement, I assume it isn't too late to provide feedback. I regret that I cannot afford to make a full review of the library that Emil deserves. So this will be just "notes". I didn't have time to study the implementation of the library. I only read most of the documentation (documentation is really good), tried some toy examples to see if I can brake it (didn't manage to), and participated in Boost discussions on the subject. Only from this am I confident that the library has been well designed. Its goals are clearly set and sound, and the designed explained in the docs. We have competing libraries in Boost for error handling (Boost.Outcome) but they put accent on different things, and besides Boost is already known to host competing designs and approaches to the same problem. The need to transport arbitrary number of information collected at different levels of call stack is a sound goal, and well addressed in LEAF. The choice to put error data on the stack is great, and well understood and explored, for instance in https://www.research.ed.ac.uk/portal/files/78829292/low_cost_deterministic_… I find function preload() (https://zajo.github.io/leaf/#tutorial-preload) really great and novel. This is the first big C++ library that I know that implements it. This really makes the error reporting code brief, clean and bug-free. Some people have reported that the usage of multiple lambdas in try-catch-like statements is inconvenient. It is; but I think it works to the advantage of the library. My approach to handling errors based on skipping dependent operations ( https://akrzemi1.wordpress.com/2019/04/25/handling-errors-is-canceling-oper…) is that you report errors (start the "stack unwinding") in many places, keep the most functions error neutral, and you stop the "stack unwinding" only in very few places. In this approach, only very few places in program code would have to put these try-catch statements with lambdas, and the clumsiness of the syntax would encourage the users to adapt the methodology "stop unwinding in very few places". My only criticism of the library is that it makes a serious design choice to use a thread-local storage and fails to communicate it clearly to potential users in the documentation. It is ok for the library to make the choice to use TLS, but it is not ok to omit it from the "design choices" in documentation. I wouldn't bring it up if the documentation was poor; but because it is so high-standard, it strikes me that this information is missing. The reservations about TLS have nothing to do with performance. First, some platforms do not provide it. But as Bjorn Reese indicated, the same functionality could be implemented with a global lock-free map with thread IDs as keys, and no TLS would be required. Second, and I think more serious, is that thread-local solutions (including the global hash map with thread IDs) play very badly with coroutines: in a coroutine world a function can start executing in one thread, and finish executing in a different one. All those TLS-based solutions compile and may pass tests but are likely to fail with random results or crashes later. Many good programmers do not realize this. Emil informs me that this problem could be worked around by manually propagating some error information at the point where coroutines are managed. And I trust him on that. But in order for the programmer to do this manual part correctly, she has to know that this problem exists, and that she has to do it. This has to be stated prominently in the documentation along with the example how to do it. This criticism is only about the documentation. Not the design or implementation. I would not use LEAF for production applications I work with. The reason is that I am satisfied c++ exceptions and I find their cost tolerable in most of the places. In the remaining parts, usually low-level, where the performance or explicit control flows are important, I do not care about the detailed information about the failure, other than an integer code. But this is just my use case. My familiarity with the domain: I report failures in my code :). I study the topic of reporting failures effectively a lot. I follow the progress of ISO C++ Committee on the subject. I have co-authored the documentation of Boost.Outcome. As I said, I didn't review the implementation, so the Review Manager will consider it in weighing my opinion, but my recommendation is to **ACCEPT** LEAF into Boost. Thank you, Emil, for writing and sharing this library. Regards, &rzej;

3 3

[review][LEAF] Review period is over
by Michael Caisse 04 Jun '20

04 Jun '20

I apologize for not sending this out earlier. The review period is officially complete. A few people have contacted me about completing reviews that they in the middle of and I will be waiting until after this coming weekend to publish the review results. Thank you to those that joined the conversation and provided a review. michael -- Michael Caisse Ciere Consulting ciere.com

1 0

[review][LEAF] Additional reviews in progress?
by Michael Caisse 03 Jun '20

03 Jun '20

Anybody in the process of producing a review to submit? Please respond here or contact me directly. -- Michael Caisse Ciere Consulting ciere.com

3 3

[review] [text] Text review begins June 11
by Glen Fernandes 03 Jun '20

03 Jun '20

The Boost formal review of Zach Laine's Unicode library, Text, commences on June 11th and concludes June 20th. We welcome the participation of Boost developers and users, and the C++ community. Quote: This library includes three layers: - The string layer, a set of types that constitute "a better std::string" - The Unicode layer, consisting of the Unicode algorithms and data - The text layer, a set of types like the string layer types, but providing transparent Unicode support Code: https://github.com/tzlaine/text Docs: https://tzlaine.github.io/text Text requires C++14 or higher, and supports GCC 5+, Clang 3.6+, and MSVC 2017. Zach initially presented the library at C++Now in 2018. He also intends to standardize a library solution for Unicode based on this work. Talk: https://www.youtube.com/watch?v=944GjKxwMBo Slides: https://tinyurl.com/boost-text-slides Try the library today, and please consider submitting a review, or even contributing to the discussion next week. Glen

1 0

Re: [boost] [review][LEAF] Review of LEAF starts today : May 22 - May 31
by Bjorn Reese 02 Jun '20

02 Jun '20

On 2020-05-22 11:31, Michael Caisse via Boost wrote: > Please provide in your review information you think is valuable to > understand your choice to ACCEPT or REJECT including LEAF as a > Boost library. Please be explicit about your decision (ACCEPT or REJECT). I vote to conditionally accept LEAF into Boost. LEAF provides a new perspective on error propagation that is worth exploring in the real world. The conditions for acceptance are: 1. Provide a lower-level API for error handling which does not necessitate a pattern matching framework. I would prefer an API closer to the variant::index() and std::get<I>() combination that enables you to write the error handling as switch statements. More rationale for this request is given under the design evaluation. > - What is your evaluation of the design? LEAF propagates error objects across multiple function calls by letting the callee store the error payload directly in a stack object, called the error context, at the caller. A pointer to the error context is located in thread-local storage. When there are more error contexts on the call stack, they are daisy chained via pointers in the child error context. This is a sound design, which gives LEAF an efficient way of propagating error payloads then a failure occurs. I like that the returned error_id can be passed through C code. There is runtime overhead in creating and destroying the error contexts even when no failure occurs. Whether or not that overhead is acceptable depends on circumstances. It seems to me that the noexcept performance overhead in the successful case and the failure case are not that different (assuming the error payload is not too heavy), so it may be a good idea to choose LEAF when worst-case performance is more important than average-case performance. The more I experimented with LEAF, the more I wished that there was an error handing API with less syntatic sugar. 1. Single-stepping into a TryBlock is really cumbersome. The developer has to single-step through layers of unfamiliar templated code before reaching their own TryBlock. This interrrupts their flow. Setting a breakpoint in the TryBlock helps but is still cumbersome. 2. Some compilation errors are difficult to interpret. For example, if you accidentally omit a required return statement in one of the try_handle_all handlers, then the compiler reports the start of the try_handle_all statement and some internal LEAF voodoo. If you are experienced with reading templates errors, then you may be able to discover the erroneous handler from the LEAF template parameters in the compiler output. The LEAF handling is reminiscient of a variant visitor, but with the variants we can use index() and std::get<T> in a switch statement instead of visitation. I would like to see something similar for LEAF error contexts. I am undecided about the reliance on thread-local storage. Initially I thought that we could just replace thread-local storage with a home- made lookup table using atomics (which are generally available on OS-less platforms, whereas mutexes etc are not,) but Niall's mention of consteval made me unsure (damn you Niall ;-) However, I do not see this as a show-stopper. The try_handle_all() and try_handle_some() functions have different return values. While I understand the rationale given in the reference documentation, it may still be confusing to users if they change from one to the other. This could possibly be solved by making it more clear in the documentation. The remote_handle_all() name is confusing. I would rename it to nested_handle_all(). > - What is your evaluation of the implementation? The implementation is of high quality, and reasonably readable if you are comfortable with templates. I did not do a proper code review, so the following are random notes. Some classes (leaf::result and the internal optional) uses placement-new and explicit destructor calls. I suggest that you create your own construct_at() and destruct_at() which uses the std counterparts when available, and placement-new / explicit destruction otherwise. This will make the code more readable. std::ostream is entangled into various classes for printing. I would prefer if printing was moved into seperate X_io.hpp files to speed up compilation times when not using them. The LEAF_CONSTEXPR should be renamed to LEAF_CXX14_CONSTEXPR to align with the BOOST_CXX14_CONSTEXPR naming. Also the #if should use '>=' instead of '>'. config.h includes <Windows.h>. This should be replaced with Boost.WinAPI. > - What is your evaluation of the documentation? Overall the documentation is well-presented. The examples reconfigures std::cout to throw exceptions. That obscures the examples because now the reader has to learn LEAF and how std::cout exceptions work at the same time. It would be better to move the std::cout exception stuff into a separate example which is only about std::cout. > - What is your evaluation of the potential usefulness of the library? LEAF presents a unique error handling solution, that may be useful to certain users. I do not expect a huge user base though, as LEAF users mainly are going to be projects with requirements not met by the alternatives. > - Did you try to use the library? With which compiler(s)? Did you > have any problems? I experimented with small examples and found no compiler-related issues. > - How much effort did you put into your evaluation? A glance? A quick > reading? In-depth study? About 6 hours. Spent most time trying understanding the design than reviewing the code and documentation. > - Are you knowledgeable about the problem domain? Yes.

6 10

Relevance of platfrom-specific shared_ptr implementations
by Peter Dimov 01 Jun '20

01 Jun '20

Greetings, shared_ptr has a number of platform-specific implementations for its control block, specifically the atomic reference count manipulation: https://github.com/boostorg/smart_ptr/blob/develop/include/boost/smart_ptr/… Those have been accumulated over the years and I'd like to clean them up, remove the unused ones and fold as many of them as possible into a few generic implementations using either std::atomic, GCC atomic intrinsics (https://gcc.gnu.org/onlinedocs/gcc-10.1.0/gcc/_005f_005fatomic-Builtins.html) or GCC/intel sync intrinsics (https://gcc.gnu.org/onlinedocs/gcc-10.1.0/gcc/_005f_005fsync-Builtins.html) But I have no experience with non-x86 platforms and don't know where boost::shared_ptr is still used and which of the platform-specific implementations are still needed. (Neither do I know on which platforms the atomic and the sync intrinsics actually work, as opposed to just compile and then fail to link.) Any information along these lines is welcome. If you know of a platform where boost::shared_ptr is still used, but which supports none of the "modern" ways to implement atomic ops, please let me know. If you would like some of the plaform-specific files to be retained, please let me know. Thanks in advance.

3 2

[boost-ml] recommended way-of-working on the ml leaks locations data.
by degski 01 Jun '20

01 Jun '20

As email clients report local time, the way we (are supposed to) post results in leaking of one's location (if one would analyze, which is easily possible with the archive), first of relative location, but once you triangulate a few (and make a few assumptions, like Niall Douglas is probably in Ireland), you can work them all out. Maybe I'm the person who finds this unpleasant, but I don't find this ideal. degski -- @systemdeg https://www.instapaper.com <http://instapaper.com> "We value your privacy, click here!" Sod off! - degski

1 0

Re: [boost] [Boost-users] [review][LEAF] Review of LEAF
by Niall Douglas 01 Jun '20

01 Jun '20

On 01/06/2020 12:51, degski via Boost-users wrote: > Is there scope for merging LEAF and Outcome as it seems from (very > detailed discussions) that they only partially overlap? My own opinion: 1. You may remember that Emil was one of two people who voted to reject Outcome back during its review, as he felt that it was the wrong design. His comments on what would be better back then map well onto what he has presented in LEAF. I find that very reasonable: (i) he thought he could do better (ii) he went ahead and did better in his opinion (iii) he is presenting now what he thinks is better, and is asking Boost for its opinion on whether he improved things or not. 2. There is long established precedence for Boost containing multiple libraries that approach a problem from multiple perspectives e.g. we currently have three error handling libraries, four maths libraries etc. If LEAF were accepted, you'd now have four instead of three error handling libraries. And that seems to me just fine, and not out of precedence. 3. I specifically and deliberately chose to not design Outcome like LEAF. I thought then, and I still do, that Outcome ought to be great at all the things no other solution is great at. I think I succeeded on that. As I already stated in my own review of LEAF, I personally think that there is already a perfectly good implementation of LEAF shipping in any C++ compiler of the past twenty years. I don't, personally speaking, see the value add for choosing LEAF over C++ exceptions except to tick some box called "C++ exceptions globally disabled compatible". > I don't think > that 'we' should keep on adding distinct libraries that partially > overlap, Existing libraries should be expanded instead, and > deduplication should be done in a coordinated way (in this case between > LEAF and Outcome). Something has to be done to stop exponential growth > of Boost while maintaining flexibility to add code at will. The best way to stop exponential growth of Boost, if that is desirable, is to start culling libraries that aren't maintained nor useful. Or at least publish a documented list of libraries that new code ought to not use. > As C++ already has exceptions (and will have exceptions light in the > future), the use-case you presented to use LEAF surgically with C-api's > seems convincing (it presents a better case than the case of generically > re-working an exception system that works, is fundamental to C++ and is > promised to be improved in the future, i.e. it's the safest bet. Any > library is secured till maintenance ceases, the language/STL > (exceptions, exceptions light) is a pdf-file (i.e. future proof by default). The current formulation of lightweight exceptions before WG21 is literally Experimental Outcome in language form. Same proposed error object, std::error. It is currently expected that both value-based and type-based exceptions would exist in future C++, in order to retain backwards compatibility, as the semantics between the two are not currently believed possible to be made exactly one-one. Experimental Outcome works fine in C code, same as how the most recent proposal for lightweight exceptions also supports C. Niall

2 3