1

Topic: Bottle neck

There is one programming language with a heap  and a functional. There is other language with a heap of libraries and a functional. In system that is used also and that language. And here necessity from one part came to cause a functional of other part. Directly it is impossible to make it. It is necessary to drag somehow all requests through one bridge. A question - how to expand a functional of this bridge without expanding the bridge? That comes to the very first mind - to make function of type Call (const char* funcName, const char* args) In the output agent  arguments and to cause that that is necessary. On idea through it it is possible to drag generally all calls. A bottleneck in  - parsing of requests. Any other decisions can be invented?

2

Re: Bottle neck

Hello, Amon_RA, you wrote: A_R> any other decisions Can be invented? RPC/RMI/COM/DBUS or other broker?

3

Re: Bottle neck

Hello, Amon_RA, you wrote: A_R> There is one programming language with a heap  and a functional. A_R> there is other language with a heap of libraries and a functional. And it is possible to sound, how these two languages are called?

4

Re: Bottle neck

Hello, Kernan, you wrote: K> Hello, Amon_RA, you wrote: A_R>> any other decisions Can be invented? K> RPC/RMI/COM/DBUS or other broker? One part - , the second part - already depends on a platform. Therefore actually and sharing.

5

Re: Bottle neck

Hello, Pzz, you wrote: Pzz> And it is possible to sound, how these two languages are called? One part -  - on With ++ the Second part - depends on a platform. On a poppy it obzh-with, on Windows it .

6

Re: Bottle neck

Hello, Amon_RA, you wrote: Pzz>> And it is possible to sound, how these two languages are called? A_R> one part -  - on With ++ A_R> the Second part - depends on a platform. On a poppy it obzh-with, on Windows it . So them all it is possible the friend from the friend directly to call.

7

Re: Bottle neck

Hello, Pzz, you wrote: Pzz> So them all it is possible the friend from the friend directly to call. And  the code turns to noodles of the conditional compilations.

8

Re: Bottle neck

Hello, Amon_RA, you wrote: Pzz>> So them all it is possible the friend from the friend directly to call. A_R> and  the code turns to noodles of the conditional compilations. From what? It is necessary to be defined simply who whom calls, and to formulate the platformenno-independent interface in the form of a dial-up  (i.e., , instead of ) functions. Also there will be no conditional compilation.

9

Re: Bottle neck

Hello, Pzz, you wrote: Pzz> It is necessary to be defined simply who whom calls, and to formulate the platformenno-independent interface in the form of a dial-up  (i.e., , instead of ) functions. .  in that and the question was - how to expand, without expanding. It is clear that through the uniform interface of dialogue it is possible to drag all, but this interface will extremely be inflated, if on each bunch separate function to get.

10

Re: Bottle neck

Hello, Amon_RA, you wrote: A_R> Phhh.  in that and the question was - how to expand, without expanding. It is clear that through the uniform interface of dialogue it is possible to drag all, but this interface will extremely be inflated, if on each bunch separate function to get. That comes to the very first mind - to make function of type Call (const char* funcName, const char* args) In the output agent  arguments and to cause that that is necessary. On idea through it it is possible to drag generally all calls. A bottleneck in  - parsing of requests. Here instead of dragging all calls through the handler, I offer  them directly. Informative functions from it any more does not become, but it will not be necessary also a branchy parcer with the manager to write. Which on the substance of absolute useful does not add anything.

11

Re: Bottle neck

Hello, Pzz, you wrote: Pzz> Hello, Amon_RA, you wrote: A_R>> Phhh.  in that and the question was - how to expand, without expanding. It is clear that through the uniform interface of dialogue it is possible to drag all, but this interface will extremely be inflated, if on each bunch separate function to get. If bunches too much - it is necessary to reconsider the interaction interface. Pzz> that comes to the very first mind - to make function of type Pzz> Call (const char* funcName, const char* args) Pzz> In the output agent  arguments and to cause that that is necessary. On idea through it it is possible to drag generally all calls. A bottleneck in  - parsing of requests. Pzz> here instead of dragging all calls through the handler, I offer  them directly. Informative functions from it any more does not become, but it will not be necessary also a branchy parcer with the manager to write. Which on the substance of absolute useful does not add anything. It agree that is necessary . Export side benefits - 1. The interaction interface is fixed and  in a separate place. 2. Changes of the interface of interaction it is supervised by the compiler from both sides.

12

Re: Bottle neck

A_R> Phhh.  in that and the question was - how to expand, without expanding. It is clear that through the uniform interface of dialogue it is possible to drag all, but this interface will extremely be inflated, Only so at you at least compilers something to check , and "with const char* funcName", if it not  for a long time settled API, I at all do not represent how to live.

13

Re: Bottle neck

A_R> In the output agent  arguments and to cause that that is necessary. On idea through it it is possible to drag generally all calls. A bottleneck in  - parsing of requests. A_R> any other decisions can be invented? Parsing of requests - how much this place is critical? It is impossible to optimize? And so, it is possible to transfer text requests and not to be soared. Full .