On Tuesday, 7 October 2014 at 01:11:42 UTC, Shammah Chancellor
wrote:
On 2014-10-06 22:28:52 +0000, Andrei Alexandrescu said:
On 10/6/14, 12:59 PM, Shammah Chancellor wrote:
I'm willing to put in the work if Walter is on board also. I
don't want
to do all that work to end up being a DPL Foundation in name
only.
That's very generous of you, thanks! We'll discuss this. From
what I read at http://lwn.net/Articles/561336/, my
understanding is that we'll need significant ongoing expenses
in additional to the initial setup expenditure of time and
money. Anyone who knows more about this please chime in. --
Andrei
I'll start doing some research. To be forward, I am 100%
ignorant of the process at the moment as well.
-S.
Very good idea. If you want i can help you.
I can suggest a few directions of to explore.
Source code toolkit: analytic, transformation...
I see this as script engine like js(dmdscript?) with extension
for access to source code in object form, like this:
void foo(int i){
int e;
if(i<0){
e = i;
}
write("yes we can!");
}
for script to be
{
functions:{
foo:{ arg: {i: {type: int, default:no }, body:[
{stament: "var_declare", name:"e", type: "int", default:
no},
{stament: "if", expression: "how it do this help!",
/* how named main code block? */ block: {}
else: no
},
{stament: "call", name:"write", arglist: ["yes we can!"]}
]
}
}
}
Yes already have similar solutions but they usually focus on a
static analysis of the code in one language.
But may be this is an expandable basis to do much more:
code analysis,
refactoring a large project simply by answering the wizard's
questions,
a project update for the new api,
and my favorite migration between languages.
Migration between languages - this i think is especially impotant
because if you have project greater than hello word you will
actually slave chosen programming language.
Let's liberate the algorithms of their actual implementation and
developers from the monotonous work!
Very wrong if you code on the 90% is a wrapper, polymorphism, or
has long been created.
