[cfe-dev] Update to C++0x feature table

[Douglas Gregor]

On Dec 20, 2010, at 9:52 PM, Michael Price - Dev wrote:

> 
> On Dec 20, 2010, at 11:46 PM, Douglas Gregor <dgregor at apple.com> wrote:
> 
>> On Dec 20, 2010, at 8:34 PM, Michael Price wrote:
>> 
>>> I've attached a second patch.  Could someone with privs commit it?  Also see inline comments below.
>>> 
>>> On Mon, Dec 20, 2010 at 5:13 PM, Alex Rosenberg <alexr at leftfield.org> wrote:
>>> On Sun, Dec 19, 2010 at 5:03 AM, Michael Price <michael.b.price.dev at gmail.com
>>>  > wrote:
>>>  > The diff is attached.
>>>  > I used the following sources:
>>>  > Bjarne - http://www2.research.att.com/~bs/C++0xFAQ.html
>>>  > Scott Meyers - http://www.aristeia.com/C++0x/C++0xFeatureAvailability.htm
>>>  > Wikipedia - http://en.wikipedia.org/wiki/C++0x
>>> 
>>> Some additional items and different [not useful] organizations of the
>>> features are here:
>>> 
>>> http://wiki.apache.org/stdcxx/C++0xCompilerSupport
>>> http://gcc.gnu.org/projects/cxx0x.html
>>> 
>>>  > Things I like about my changes:
>>>  > 1. Near complete listing of core language changes.
>>>  > 2. Links to STL implementation status pages.
>>>  > 3. I like the classification of features.
>>>  > Things I don't like about my changes:
>>>  > 1. I would love to provide links into the current draft for each
>>> feature,
>>>  > but I couldn't find a way to link into the draft PDF.
>>>  > 2. I wish there were a better way to classify features (as some are
>>>  > cross-cutting).
>>>  > Indifferent:
>>>  > 1. Draft proposal documents.  I included and expanded them because
>>> they were
>>>  > there before.
>>>  >
>>>  > There is still one extra thing that I haven't done, which is to
>>> provide some
>>>  > sort of desired priority, or if that is too difficult, some way to
>>> indicate
>>>  > dependent relationships between features.
>>> 
>>> A few paper numbers to add:
>>> * sizeof on members without object instance is N2150, 5.1.1p10
>>> * nullptr is also N2214
>>> * char16_t and char32_t are part of N2249
>>> 
>>> Other than that, it would be most helpful if current status were
>>> filled in so we all know what needs to be done.
>>> 
>>> Agreed.  Alas, I don't know enough to handle that, except for perhaps the type deduction stuff.
>>>  
>>> In my ideal world, as
>>> the status in each box is changed, a revision number is put in there
>>> since "releases" are infrequent.
>>> 
>>> 
>>> I had the same thought originally, but wasn't sure how to fit that information in with everything else.  I decided to separate the explanation of test status to another paragraph, and just stick an "rXXXXXX" in the complete block.  I figure that people could discern what that was for when they saw it.
>>>  
>>> Somebody more familiar with the draft should comment if any items are
>>> missing.
>>> 
>>> 
>>> There are sure to be some things I missed... anyone else looking?
>> 
>> This is great, Michael, thanks!
>> 
>> I made a few edits, dropping long double (which was part of C++98/03) and new function declarator syntax (which isn't in C++0x), updating status, etc., and committed as r122315.
>> 
>> 	- Doug
> 
> By new function declaration, I was meaning the trailing return type as in:
> 
> template <typename T, typename U>
> auto func (T t, U u) -> decltype(t*u);

Ah, I see. I've put back that row, but with only the link to the late-specified return type paper. Thanks!

	- Doug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20101220/0adba27f/attachment.html>