On Friday In Madrid, Spain, the ISO C++ committee approved the Final Draft International Standard (FDIS) for C++0x. So C++0x is now complete in a technical sense. There is no possibility for further technical change. It will be a few weeks before the actual FDIS document is available. The ISO formal voting process will take until August or later. It's an up or down, yes or no, vote. All the countries actively participating voted yes for shipping the FDIS, so this ISO vote is mostly a formality. Most of the changes approved in Madrid were minor wording changes fixing mistakes in specification. Proposals to remove several features on the grounds they haven't been implemented or tested were rejected, usually by overwhelming vote margins. For example, the range-based-for-loop problem that surfaced on the Boost list caused a tweak in name lookup wording. On the library side, a lot of effort was devoted to "noexcept". A much more coherent set of usage guidelines was agreed on, and that meant going through the standard library and removing a lot of uses of "noexcept" that were over aggressive. More on that later, since Boost will probably want to use a similar set of guidelines. A straw poll for restarting work on Library Technical Report 2 passed without opposition. There was a lot of private discussion as to which Boost libraries committee members wanted to see go into TR2. --Beman
On 3/27/2011 9:50 AM, Beman Dawes wrote:
On Friday In Madrid, Spain, the ISO C++ committee approved the Final Draft International Standard (FDIS) for C++0x.
So C++0x is now complete in a technical sense. There is no possibility for further technical change. It will be a few weeks before the actual FDIS document is available.
The ISO formal voting process will take until August or later. It's an up or down, yes or no, vote. All the countries actively participating voted yes for shipping the FDIS, so this ISO vote is mostly a formality.
Most of the changes approved in Madrid were minor wording changes fixing mistakes in specification. Proposals to remove several features on the grounds they haven't been implemented or tested were rejected, usually by overwhelming vote margins. For example, the range-based-for-loop problem that surfaced on the Boost list caused a tweak in name lookup wording.
On the library side, a lot of effort was devoted to "noexcept". A much more coherent set of usage guidelines was agreed on, and that meant going through the standard library and removing a lot of uses of "noexcept" that were over aggressive. More on that later, since Boost will probably want to use a similar set of guidelines.
A straw poll for restarting work on Library Technical Report 2 passed without opposition. There was a lot of private discussion as to which Boost libraries committee members wanted to see go into TR2.
Needless to say, thanks to everyone on the C++ standards committee for the work they have done. Although C++ programmers may disagree on decisions by the committee, there can be no doubt that the people on the committee do their work without profit and at considerable personal expense, for the benefit of millions of C++ programmers throughout the world, with little real appreciation for the efforts they provide. It would be great, once the final standard is approved, if you, or anyone else on the C++ standards committee, can send out a message on the main mailing lists and NGs about where/how an individual programmer can obtain a copy of the final C++ standard in all its glory.
participants (2)
-
Beman Dawes -
Edward Diener