stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <>
Subject Re: New 27.basic.ios.cpp test migrated (LONG)
Date Wed, 16 Apr 2008 00:32:49 GMT
Eric Lemings wrote:
>> -----Original Message-----
>> From: Martin Sebor [] 
>> Sent: Tuesday, April 15, 2008 4:26 PM
>> To:
>> Subject: Re: New 27.basic.ios.cpp test migrated (LONG)
> ...
>> About the test, it looks good to me. It needs only a few minor
>> tweaks :)
>> Regarding the convention for preprocessor directives, it is
>> 2 spaces and if there are places where it's inconsistent they
>> should be fixed :)
> Is there a reason its 2 spaces instead of 4?  If not, it should
> be fixed everywhere i.e. made consistent in all contexts. :)

Not one I know of. It's just something we inherited/adopted over

> Hmm.  I noticed variable names for command-line options in other
> tests using the rw_ prefix so I used it also.  Funny how undesired
> conventions get propagated, huh?

Some of them probably crept in from the driver.

> Good eye.  Is there a rationale for this or is just preference?
> For me, this is just a holdover from K&R C.

Same as with the preprocessor directives.

> Also, I just noticed the link to the style document on the home
> page is broken.  Is it stored in Subversion or some other place
> maybe?  Probably should give it a quick glance...

It's not broken -- it's a placeholder for a future document. Feel
free to start a page documenting the existing style on the Wiki.

>> 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).
> Yep, and return types of functions default(ed?) to `int' type when
> omitted though I'm not sure that still holds but they're still
> always explictly specified nonetheless.  Same premise: never
> rely on the default.

C++ doesn't allow missing return types. I don't think C99 does


View raw message