• berni (4/4) Aug 25 2019 Out of curiosity: Browsing the source of stdio.d I found that
• Jonathan M Davis (13/17) Aug 26 2019 The dot makes it so that it's specifically referencing a module-level sy...
• a11e99z (19/34) Aug 26 2019 OFFTOPIC:
• berni (2/5) Aug 26 2019 Ah, thanks. Now it makes sense! :)

berni <someone somewhere.com> writes:

Out of curiosity: Browsing the source of stdio.d I found that 
flush() is implemented by calling fflush from some C++ library. 
What I don't understand: Why is the call to fflush preceded by a 
dot?

Jonathan M Davis <newsgroup.d jmdavisprog.com> writes:

On Sunday, August 25, 2019 11:59:08 PM MDT berni via Digitalmars-d-learn 
wrote:
 Out of curiosity: Browsing the source of stdio.d I found that
 flush() is implemented by calling fflush from some C++ library.
 What I don't understand: Why is the call to fflush preceded by a
 dot?

The dot makes it so that it's specifically referencing a module-level symbol
(be it in that module or an imported module) instead of a local or member
symbol.

https://dlang.org/spec/module.html#module_scope_operators

In this particular case, it doesn't look like the dot is necessary, because
File doesn't have an fflush member, but there are a number of cases where it
has a member that's the same as a C function that it wraps, in which case
the dot would be necessary to reference the C function instead of the member
function. So, it wouldn't surprise me if whoever wrote that code was just
putting a dot in front of C function calls in general.

- Jonathan M Davis

a11e99z <black80 bk.ru> writes:

On Monday, 26 August 2019 at 09:14:23 UTC, Jonathan M Davis wrote:
 On Sunday, August 25, 2019 11:59:08 PM MDT berni via

 - Jonathan M Davis


OFFTOPIC:

(dont have ur email. dont like emails cuz too officially and too 
long)
(and dont want create new topic. this one probably 
solved/finished already)

about benchmark 
https://dlang.org/library/std/datetime/stopwatch/benchmark.html
idk all reason why that not vice versa but imo better to change:

for now:
Duration[3] benchmark!(f1, f2, f3)( int runCount );

remarks:

 benchmark!(
         // next is simple func list
 	() => ps.each!( p => b += gcd( p[0], p[1])),
 	() => ps.each!( p => a += divGCD( p[0], p[1])),

 )( 10 ) // and here real call with complicated number

 	.array // without it .map! is not compiling
 	.map!( x => x.total!"msecs")
 	.writeln;

1) .map! for static arrays doesn't compiling
   maybe better to return Range?

2) code with benchmark!() looks turn upside down when u try to 
use lambdas as args
I suggest add another versions of benchmark:
auto bench!( int N, FuncList...)( FuncList funcs );
 benchmark!10( () => ps.each!( p => b += gcd( p[0], p[1])),
 	      () => ps.each!( p => a += divGCD( p[0], p[1])))...

auto bench!( FuncList...)( int N, FuncList funcs );
 benchmark( 10,
 	() => ps.each!( p => b += gcd( p[0], p[1])),
 	() => ps.each!( p => a += divGCD( p[0], p[1])))...

berni <someone somewhere.com> writes:

On Monday, 26 August 2019 at 09:14:23 UTC, Jonathan M Davis wrote:
 The dot makes it so that it's specifically referencing a 
 module-level symbol (be it in that module or an imported 
 module) instead of a local or member symbol.

Ah, thanks. Now it makes sense! :)