incubator-stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <se...@roguewave.com>
Subject Re: New 27.basic.ios.cpp test migrated (LONG)
Date Wed, 16 Apr 2008 16:31:27 GMT
Eric Lemings wrote:
>  
> 
>> -----Original Message-----
>> From: Eric Lemings [mailto:Eric.Lemings@roguewave.com] 
>> Sent: Wednesday, April 16, 2008 9:39 AM
>> To: dev@stdcxx.apache.org
>> Subject: RE: New 27.basic.ios.cpp test migrated (LONG)
>>
>>  
>>
>>> -----Original Message-----
>>> From: Martin Sebor [mailto:msebor@gmail.com] On Behalf Of 
>> Martin Sebor
>>> Sent: Tuesday, April 15, 2008 10:46 PM
>>> To: dev@stdcxx.apache.org
>>> Subject: Re: New 27.basic.ios.cpp test migrated (LONG)
>>>
>>> Martin Sebor wrote:
>>> [...]
>>>> Finally, every function is extern by default. There is no need
>>>> to explicitly declare it as such (I believe there are compilers
>>>> that warn about function definitions with the extern keyword).
>>> FYI, here's an example of the diagnostic (issued by HP aCC 6)
>>> from our nightly builds I was referring to above:
>>>
>>> "$(TOPDIR)/src/ti_num_get.cpp", line 51: remark #4244-D: 
>>> extern storage 
>>> class used with a function definition
>>>    _RWSTD_DEFINE_FACET_FACTORY (extern _RWSTD_EXPORT, 
>>> num_get, TARGS_C, 
>>> num_get);
>>>    ^
>> Another "useful" compiler warning?  I'm not surprised.  HP aCC 6 has
>> about as many real warnings as MSVC 8 and 9.  You just can't put much
>> stock in compiler warnings these days.
>>
>> I wonder what the rationale for the warning is?  The syntax and
>> semantics
>> of C and C++ clearly allows extern specifiers in function definitions.
>> I'll have to look it up in HP's aCC docs.
> 
> Yeah, here's the "rationale" for the warning:
> 
> http://www.docs.hp.com/en/11925/Diagnostics_Documentation/Diagnostics_in
> dex.htm#4244

Good link!

> 
> In short, there is no reason for the warning; it's just more nonsensical
> verbage.

Just like defining main() extern is unnecessary verbiage.

The point is that every remark, however useless, contributes to
the noise from our builds and to the amount of time some of us
have to spend poring over the build logs trying to find those
that might indicate a problem. So let's please try to avoid
introducing new ones if we can help it.

Martin

Mime
View raw message