WARNING: This server is unstable and will be retired in the next days. If you want to keep this forum available, please request immediately a migration on the Nabble Support forum. Forums that don't receive any migration request will be deleted forever.

 « Return to Thread: CInt and function pointers...

Re: C sharp script...

by gotom :: Rate this Message:

| View in Thread

Hello Tarmo,

It's a great idea to start from C# and try to replace C++ interpreter with better implementation.
Several people have tried similar thing in past 15 years. The biggest challenge was C++ complexity. 
No one has completed the work, to my knowledge.  
There is something going on in CERN to replace CINT.  They may have power to accomplish it.

I'd like to state fundamental benefit of C++ interpreter like CINT.   Key thing is that we can write
performance hungry code and script in same language.  This is crucial for people who does serious
simulation , data analysis , etc.. requiring the best-of-the-date performance.  I know such population is 
very small, but important.    

Good luck
Masaharu Goto

On 2012/04/11, at 5:52, Tarmo Pikaro wrote:

Hi !
 
I have tried also C# script engine - http://www.csscript.net/ .
 
Error handling is not perfect one, but at least it looks like working - at least with some simple C# scripts.
(Managed to parse xml, etc...)
 
Also tried Cint with newest Autocad object arx include headers - and makecint refuses to swallow all
new and fancy C++ templates.
 
In programming language wise also it's easier to start developing from clearer implementation
than from cryptic one - compare code of cint versus c sharp script.
 
So I'm still thinking about may be starting from C#, and then returning back to C++ if needed,
but I suspect that C# memory fingerprint is not so good... (#1)
 
Finally I have found similar kind of data types as in perl - perl operates in scalar, array, hash terms
(and any complex permutation out of it) - in C# there exists similar kind of classes like ArrayList and
Hashtable. Still missing scalar.
 
I guess most intrested feature for makecint - is to be able to remote API by providing necessary clue
- .h / .lib for C++ objects.
 
If I would start writing new source code parser, I would write it's scratch in perl.
For final code - C# or C++. (#2)
 
Based on ideas #1 & #2 above - may be we could start from basic data types (or objects ?) -
I would like to drag perl's data types into C++ / C#. Parse C++ / C# using basic data types, construct
abstract syntax tree of source code, and then transform one programming language to another.
 
It's quite obvious that eventually switching programming language could be good alternative,
we could start to identify base concept of how such language switch could be performed.
 
Would you prefer to chat over skype ?
 
--
Have a nice day!
Tarmo.

 « Return to Thread: CInt and function pointers...