1

Topic: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: Those> Somebody else uses async io? In favor libuv says that it is superficial a C library, it use all who feel like it, well the vivid example node. And async io it is fastened on a C ++ and the sense of its usage outside of a C ++ is not clear. Yes.

2

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: Those> Somebody else uses async io? How it is possible not to use ? If only in  hand-made articles. A question that in what? Both that and another quite apprx.

3

Re: Re: libuv vs boost:: async-io

In what a question? Both that and another  is popular in different projects, descend on  a site libuv for example https://groups.google.com/forum/#!forum/libuv

4

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: Those> Somebody else uses async io? The sense in asynchronous input-output appears in that case when an amount of simultaneous operations essentially exceeds number of kernels of the processor. If such operations (roughly speaking to connections if to speak for example about a network) a little normal synchronous (locking) operations will be more effective (by the way in Boost. Asio they too are available). Those> In favor libuv says that it is superficial a C library, it use all who feel like it, well the vivid example node. And async io it is fastened on a C ++ and the sense of its usage outside of a C ++ is not clear. Well owing to different given API libuv it is accessible from the majority of languages, and Boost. Asio only from a C ++ (but therefrom it is as much as possible convenient). So authors also reflected. And in what a question that?)

5

Re: Re: libuv vs boost:: async-io

Hello, alex_public, you wrote: Those>> In favor libuv says that it is superficial a C library, it use all who feel like it, well the vivid example node. And async io it is fastened on a C ++ and the sense of its usage outside of a C ++ is not clear. _> well owing to different given API libuv it is accessible from the majority of languages, and Boost. Asio only from a C ++ (but therefrom it is as much as possible convenient). So authors also reflected. And in what a question that?) in that, certainly, that the author does not know Vpros, what semantics, according to the modern trends, to give here to it: "outside of a C ++". He would like to listen to judgements of colleagues,  most. Such statement of "question" means some internal struggle which began explicitly not yesterday.))

6

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: _>> Well owing to different given API libuv it is accessible from the majority of languages, and Boost. Asio only from a C ++ (but therefrom it is as much as possible convenient). So authors also reflected. And in what a question that?) those> Really at someone arise difficulties to call C - function from a C ++? To cause - any problems is not present. But  OOP is always more convenient .

7

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: _>> to Cause - no problems are present. But  OOP is always more convenient . Those> So the C ++ under a cowl also causes  methods,  OOP is finite. But all these uglinesses happen in the library code, instead of in your application-oriented code.

8

Re: Re: libuv vs boost:: async-io

Better it is finite when it happens in the space, instead of I walk through a heap of other libraries anyway more for such approach superfluous essence - only worse and the rector not such a difficult pattern is easier libuv or asio it turns out, hardly hardly the code round several

9

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: _>> to Cause - no problems are present. But  OOP is always more convenient . Those> So the C ++ under a cowl also causes  methods,  OOP Not absolutely. With ++ gives freedom on types, and With beats specific types and algorithms nails to each other. Implementation libuv it is far not most  and often not approaching on this reason. Not effective including, because on With.

10

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: V>> Not effective including because on S.Tyo> Suddenly on a C ++ it is heavier than the program and , than on a C. Different, it is impossible to compare programs directly. In basis of library With ++ give more effective code, than S.Bolee's similar libraries  they it is normal from  multifunctions and  reliabilities. Those> Emulation OOP . Where also what emulation?)) matter is not in With or With ++. Was specific about libuv the known rule works - the common decision is the worst for a specific case. If it would be desirable an efficiency maximum any libuv it is not necessary. Corresponding  OS is already quite sufficient for needs of the programmer of "library".))

11

Re: Re: libuv vs boost:: async-io

Miles sorry, and in what sphere and in what language you would program what to do such strange outputs? More often all on the contrary

12

Re: Re: libuv vs boost:: async-io

Hello, Tyomchik, you wrote: Those> Somebody else uses async io? In favor libuv says that it is superficial a C library, it use all who feel like it, well the vivid example node. And async io it is fastened on a C ++ and the sense of its usage outside of a C ++ is not clear. At me in sciter was  analog asio still 10 years ago. When decided this part to rewrite (for the several reasons) that selected between libuv and asio As a result stopped on libuv. The Principal cause - more developed functionality. As an example: pipes (named pipes / local sockets), child process c interception possibility stdin/stdout and still it is a lot of that from assortment node.js. Actually Sciter now (in standard delivery) it HTML/CSS/script + something of type NodeJS on functionality. That that is not pleasant in libuv - the strange policy of possession of objects. Instead of in general customary reference counting primitives there is something similar, but "idempotent": void uv_ref (uv_handle_t* handle) Reference the given handle. References are idempotent, that is, if a handle is already referenced calling this function again will have no effect. uv_ref (handle); uv_ref (handle);... uv_unref (handle);//here it finishes operation of that handle. Well i.e. there not refcounter, and simply certain bool - referenced or not.

13

Re: Re: libuv vs boost:: async-io

All by hands it is written on  epoll or kqueue for an evening-two. Through a week of the free evenings at you on hand ideal  for , gathering ideally under two your platforms. Windows, naturally,  for it is not necessary. Well if the Windows so are necessary, still to steam of evenings, steam of macroes ifdef in the  and it  at you also .