dumblob

Why don't you want to use (class) decorators for such repetitive patterns?

All things said. @Night-walker precisely described our perception and the situation. Now we have to wait.

To bring up a real scenario: ``` ruby load web.json y = { (enum)$a -> 5, $b -> 7, } io.writeln( json.serialize(y) { [obj] switch (obj) type { #case map:...

> I guess it's needless since the type of the returned value is always known at compile time and there is no problem in specifying it directly. Yes, it's known...

`-O3` is never used in production environment (not even ATLAS uses it), so manual marking makes sense. > Also, it might be unsuitable for exported functions because it modifies the...

> Why? Because it's dangerous (e.g. programmer-optimized multithreaded code with volatile and static variables is a nightmare) and sometimes considerably slower than `-O2` (e.g. the mentioned ATLAS). > What is...

Actually [not](https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html). E.g. inlining is in `-O3` and function attributes exist (e.g. `fastcall`), because compiler has no way how to determine which function does what (and therefore it's not part...

> `-O2` implies `-finline-small-functions`. Yes, small functions, but not functions called frequently => function attributes added manually. > If it's that simple, I guess the compilers should again handle that...

Yep, I like DPDK :) But as you might have noticed, it's mostly about shuffling with fixed-sized data and doing arithmetic calculations. Nothing less, nothing more. I doubt function attributes...

> I don't think we can gain much performance improvements by tuning compiling flags I think it could be about 5-10%. > the main overheads are in the following places:...