FeedaMail: Comments for Sutter’s Mill

	Comments for Sutter's Mill
	(Sponsor this newsletter) Comment on Trip report: Winter ISO C++ standards meeting by petke Thanks for the trip report. I have been refreshing this page for a week in anticipation :) What is the status of the “Concurrency 1 TS” though? It seems to me some features there are low risk, widely useful, and has a number of mature implementations. I’m thinking of the std::future improvements. Especially the “then” continuations and the heterogeneous “when_any” would make std::future much more useful. Are these not mature enough to be split out of the TS and be shipped in C++17? There are many years of of experience with using them in PPL (+power pack), TBB, HPX, etc. Is there any issues holding this back for another 3 years? Read More » Comment on Trip report: Winter ISO C++ standards meeting by jmckesson I’m glad to see many of the Library TS features go into the standard, as well as the FileSystem TS. My biggest question is about system’s integration: what is being done to actually incorporate these new library elements within the rest of the standard library? For example, FileSystem TS defines a path type. But *nothing* actually uses it directly. Not `std::fstream`, not `fopen`, not one single standard library file processing function takes it. Will `string_view` overloads be added for `std::stoi` and its ilk? And what of combinations of such elements? Will `path` allow you to get a `string_view` of its contents? Will `path` be constructible from a `basic_string_view`, just as it is from a `basic_string`? These TS’s are all designed to be stand-alone. But once they become a part of the standard, then they need to work well with each other. We really need different parts of the standard library to work well together. Read More » Comment on Trip report: Winter ISO C++ standards meeting by F.Mehrabi About Order of Evaluation: Are we going to have expressions evaluated regarding associativity and precedence rules of combining operands with operators (because that is IMHO the naturally expected behavior)? regards, FM. Read More » Comment on Trip report: Winter ISO C++ standards meeting by Herb Sutter @petke: Concurrency 1 was only just completed at the last meeting, so it was a little too early to propose merging it into trunk, as the trunk is close to shutting down to stabilize and ship C++17. @jmckesson: I believe people are busily writing papers on how to apply string_view throughout the standard library. The committee felt that shouldn’t be gating factor to ship string_view itself, which is well understood and can deliver value to user code today even as the standard library works on potentially adding it pervasively as a separate step. @farid: That’s the basic idea, in particular to have left-to-right evaluation for everything except assignment which naturally flows right to left. Check out the link to the paper for details. Read More » Comment on Trip report: Winter ISO C++ standards meeting by Ilya @jmckesson Looking at the linked (pre-meeting) versions (I have no idea what if anything have changed while merging to the standard): 1. `path` defines implicit conversion to `string` and thus can be used in every case where string is currently used, including fstream constructors. (The corresponding note mentions exactly this use case). 2. There is not conversion from `path` to `string_view`. 3. `path` already has a tepmlated constructor from a range, which accepts a string_view. Read More »
	(Sponsor this newsletter)
	Delievered to you by Feedamail. Unsubscribe