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: Optimzed vs. readable code


Howard Hinnant <hhinnant@apple.com> writes:

| On Nov 30, 2005, at 1:05 PM, Gabriel Dos Reis wrote:
| 
| > If you really want to make the codes readable, here is a better and
| > more useful (in my mind) project for you:  Implement the #nospam CPP
| > directives so that we don't have to uglify most of the names.
| 
| I would dearly like to see this done too.
| 
| Two (probably stupid) questions:
| 
| 1.  Is #nospam slang for #scope in N1614, or is it something similar
| but different?

Yes.

| 2.  Do we have a standards leg to stand on here?  I truly hope
| something like #scope makes it into C++0X.

I'll take care of it, once concepts are on their way, but I
don't mind bringing forward again before the committee in Berlin. 

| What is our current
| policy regarding really useful but not-yet-standard functionality
| like this?  I'm asking in the context of your recent (and reasonable)
| posting regarding move semantics.

I think the general policy is to do the experiments on branches.  It
mostly depends on the urgency/severity and the kind of problem we
solve, judging from past practice.  But, I guess #nospam solves a
serious problem that many library writers would like to see solved.  I
remember having had this discussion with Matt and he agreed he would
like to see a GCC implementation too -- regardless of what the committee
does later. 

At any rate, it is something we should try on a branch.  I cannot
promise anything (contrary to what I told Matt a year or so ago)
because I have my hands in too many things.

-- Gaby


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