This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Bits Question


On Sun, 2010-01-03 at 14:27 +0000, Jonathan Wakely wrote:

> libstdc++ *is* the GNU C++ standard library. No other implementation
> uses that name.
> You miread Paolo's reply, the adjective "generic" was referring to the
> mailing list, not the project.

So we agree I asked in the right place.

> > (4) You've made it apparent several times now that you do not know
> the
> > answer to my question, and repeated attempts to remind us are not
> > productive. Please do not respond to the thread and let's wait for
> > someone who has knowledge on this subject to address it.
> 
> Did you miss where he said "bits" means "little pieces"?  It doesn't
> stand for anything, it's an English word. My dictionary defines bit as
> "a small piece, part or amount of something." That's exactly how it's
> used in the context of libstdc++ directory naming.

My god, I *understand* that. I reasoned that, given the context, the
developers probably had something more technical in mind than Binary
digIT or "bits of information", such as "binary interfaces" for
architectural specific optimizations (e.g. SIMD, long mov instructions,
etc. on i686, armel, whatever). Given the context, this wasn't
unreasonable to suspect. I looked through the GNU Coding Standards and
found nothing, but found similar instances of the dir in other complex
projects that can be dependent on hardware specific features (e.g. GNU
Hurd).

> A developer working on their own project can use whatever names they
> want. If they want to put implementation details in a separate
> sub-directory they could call it "detail" or "bits" or "fred" and it's
> not relevant to the choice of names used by the libstdc++ project, and
> it's off topic for this mailing list.

It's not relevant to ISO C++ specification, but was an implementation
specific detail for GNU libstdc++'s version - hence why I asked here.

> I suggest you stop making yourself look sily and listen to people with
> knowledge of the libstdc++ project, and knowledge of what is on topic
> on this mailing list (Paolo actually told you this was the wrong list
> in his first reply, not after you found that reply lacking.)
> 
> Jonathan

You're right, I am making myself look silly continuing this dialogue
with you.

-- 
Kip Warner -- Software Developer
President & CEO Kshatra Corp.
OpenPGP encrypted/signed mail preferred
http://www.thevertigo.com

Attachment: signature.asc
Description: This is a digitally signed message part


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]