r209869 - PR11410 - Confusing diagnostic when trailing array element tries to call deleted default constructor

[Alp Toker]

On 31/05/2014 01:10, Richard Smith wrote:
> On Thu, May 29, 2014 at 8:13 PM, Nikola Smiljanic <popizdeh at gmail.com 
> <mailto:popizdeh at gmail.com>> wrote:
>
>         That should read "default-constructed".
>
>
>     Could you elaborate? English is not my native language and
>     grepping the source code didn't produce anything useful.
>
>
> I think the hyphen should only be present when "default constructed" 
> is used as an adjective. Here, it's being used as a compound verb, so 
> I think it should not be hyphenated.

"Compound verbs are either hyphenated or appear as one word. If you do 
not find the verb in the dictionary, hyphenate it."

http://www.grammarbook.com/punctuation/hyphens.asp

Alp.


>         The message can be made shorter by dropping the first half
>         without losing value -- the source location is sufficiently
>         informative.
>
>
>     I don't think it's obvious without the first part but I'll change
>     it if you insist? I've just noticed that 'omitted element'
>     probably needs a plural version.
>
>         It's unconventional to use 'was' when describing semantic
>         analysis results.
>
>
>     How about 'parameter 0% was not declared' or 'unnamed type used in
>     template argument was declared here'. There are many others and to
>     me they feel more natural, but again I'm not a native speaker. I
>     actually don't like that terse mechanical voice compilers often have.
>
>
> Nonetheless, we should use a consistent voice throughout all our 
> diagnostics.
>
> How about turning this note into a context note (which is what it 
> really is):
>
> "in implicit default construction of element with omitted initializer"

-- 
http://www.nuanti.com
the browser experts