[Ecls-list] ECL as an embedded language on an Android game

[Juan Jose Garcia-Ripoll]

On Thu, Jul 11, 2013 at 3:31 PM, Anton Vodonosov <avodonosov at yandex.ru>wrote:

> Maybe it is impossible to give a 100% guarantee to cross compile any CL
> code.
> But I guess more than 93% of useful libraries are not affected but such
> issues and
> can be cross-compiled as is, without even tuning them.
>

You know, as well as I do, that failure of a couple of libraries can ruin
quite many other dependencies :-)


> I am curious to know what are the specialized arrays? (used in clx)?
>

Buffers, prebuilt data, etc.

Note, that the target has the same runtime library - ECL.
> So the most troublesome things for cross compilation are properties
> of the underlying platform that are visible through the CL;
> not the properties of the given CL implementation.
>

I agree, but they are there.


> And the underlying platform is abstracted very much by CL.
> Endiannes is abstracted by function cl:ldb - zero byte is the less
> significant
> byte of an integer, doesn't matter how the integer layed out in memory.
>

Not really. There is code out there that uses endianness to write down data
from 32-bit buffers in one way or another. How binary files are written
depends on platform dependent details. Similar to the built int type
classes, such as integer types or specialized arrays.


> So I still think cross-compilation of Lisp code has significant potential.
>

I do not deny it. I just simply say that lisp coders are not used to write
code that works that way-


> Mocl is even trying to make business on it.
>

And they are forced to cooperate with library authors revising what they do.


> On the other hand, it is not always necessary. Loading the system
> from source code on target and running interpretter may be OK for many
> applications.


Totally agree. We are saying the same thing, just exposing different
aspects of the same problem.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20130711/c48a3fef/attachment.html>