World: r4wp

Hi guys - I am trying out Red/System and all seems very good until 
I try out Kaj's fibonacci and mandelbrot functions. I downloaded 
the c-library -sorted the paths and locations and tested the library 
on simple things and all compiles well. When I try to compile fib 
and mandel I get the following compilation error:

>> do/args %rsc.r "%tests/mine/fibonacci.reds"
Script: "Red/System compiler wrapper" (none)

-= Red/System Compiler =-
Compiling tests/mine/fibonacci.reds ...
Script: "Red/System IA-32 code emitter" (none)
*** Compilation Error: alias name already defined as: [str
uct! [
        item [c-string!]
    ]]
*** in file: %runtime/common.reds
*** at line: 61
*** near: [
    str-array!: alias struct! [
        item [c-string!]
    ]
    typed-value!: alias struct! [
        type [integer!]
        value [integer!]
        _padding [integer!]
    ]
]

Any suggestions as to what I am doing wrong?

Most probably the binding was not tested for quite some time, and 
Red + Red/System evolves, so there might be some regression there. 
Better wait for experts, though ...

Ken: thanks for giving it a try. We might have left these two scripts 
lagging behind. Anyway, you'll have that fixed during the day (the 
good side of having a fast pace). ;-)

Go Doc, Go! :-)

Thanks guys, for the speedy replies

Doc, I think the work you are doing is superb and Red is a very exciting 
project - thankyou!

I test the bindings and examples regularly, so they should work. 
I'll check, but which version of Red/System are you using?

I am using the version downloaded from Github a couple of days ago 
"v0.2.6-39-gfc5c4a3"

Both the Fibonacci and Mandelbrot examples work with the current 
dyn-lib-emitter branch of Red/System, and I tested them for the 0.2.6 
and 0.2.5 releases, so they should work there, too

Did you install the extra Red-common Fossil repository that the C 
library binding now needs? (See the Announce group here)

Ok - I downloaded the examples by copy and paste from the examples 
in the Files section of the C Library Binding for Red - is this the 
correct way to access the examples as well as the c-library or is 
there another way?

I am running on Windows 8 64 bit by the way

Currently, the Red-common repository goes into a common/ subdirectory 
but the Red-C-library repository goes into a Red-C-library/ subdirectory

It's much easier to download a repository as a Zip package or tarball. 
Go to a particular version or check-in and use the download links 
there

It's even better to check out the repositories with Fossil, so you 
can keep them up to date easily:

http://rebol.esperconsultancy.nl/documentation/how-to-use-Fossil.html

Ok - I am confused - are we talking about getting Red/System now, 
or your c-library bindings and examples? I downloaded Red/System 
from Github, the zipped version using the supplied button. Then when 
I had all that working I went to the Red blog site - contributions 
- c library binding - and downloaded your c-library etc from there 
using copy and paste

No, my bindings

I will have a look into using Fossil and try again - it may be that 
the examples I have obtained are out of sync with the examples obtained 
through fossil - once I have figured it out I will try again and 
let you know my results -thanks for your help

There is little opportunity for being out of sync, as I tested them 
all the time

Most likely, you have put files in different places then they're 
supposed to be, and the paths are now wrong

Possibly, but I did go through all the #includes and made sure the 
c-library and common.reds was accessible and working, but like I 
said, I will try using Fossil and see what happens that way

More information on handling the repositories for one binding is 
here:

http://rebol.esperconsultancy.nl/extensions/0MQ/

Thanks Kaj

You're welcome

I'm repeating myself, but it looks I'm not alone who find it confusing 
to work with 2 repos. As we are really working with just a few kB, 
it would be really better to have it in one place = GitHub and not 
to merge libs with Fosill. But anyway, I appricaiate your contributions 
Kaj, so use whatever you like.

The situation at the moment certainly can be confusing, but the solution 
lies not in moving everything to a central repository.


The real solution is to improve Red(/System)'s mechanisms for locating 
and including other scripts ("libraries"). A simple "search path" 
would already go a long way, and will probably be added soon.

(Please note that I'm deliberately using rather vague terms, as I 
only want to point at where the solution lies, but not suggest a 
concrete design of that solution, at the moment.)

it would be nice to formalise, e.g. to have /bindings subdir, where 
partricular systems would be represented by subdirs, etc.

Agreed. My bindings already use that structure among each other

They mirror the situation in the open source world, where work is 
divided in independent projects. Even if you would heap the bindings 
together, you would still have to manage the libraries separately

A distributed source control database is essential to my operations, 
so I can't just use anything that others happen to use

What's wrong with Git for your purposes?

Kaj, downloaded the Zip file from Fossil and re-ran (after sorting 
paths to files) and get the same problem as already posted. The problem 
is not file locations as that results in a different compiler error 
which is easily sorted. It seems to me in the problem I have posted, 
the compiler thinks there is already an alias called str-array! - 
however I have searched every reds file and there is no other with 
that name! Anyway, it is not urgent but I just wanted to feed back 
my findings to you.

Git is bloated, to the point that in the past we haven't been able 
to port it to Syllable. Maybe the core would be portable now, but 
we don't want it, because it's huge and complex and confusing to 
users. Fossil is much simpler

Ken, I'm not sure what revision of 0.2.6 you used. When downloading 
from Git, the "master" branch should be used. I'm also working with 
the v0.3.0 development branch, so I'll reinstall that later today 
to test

There have been many bugs in the development branches lately, so 
if you're not using the latest, all bets are off

ok - I will download v0.3.0 and try it on that too

What do you mean by sorting paths to files? If you unpack the repositories 
in the right places, no files need to be edited

Under windows there is no common folder where the common.reds file 
resides - on windows common.reds resides in the runtime folder, so 
the #include file in the c-library.reds file needs to point to where 
that file resides. Likewise on fib and mand - they point to the c-library 
which I put into the library folder  that exists in the Red/System 
folder on the windows version.

But the file locations are not the issue - the compiler complains 
accurately when the files are not in the locations specified by the 
#include directive.

Ah, see, you're mixing up files. The common.reds file in Red has 
nothing to do with my bindings. It also has nothing to do with Windows. 
You should leave that alone and put my bindings in the subdirectories 
I mentioned in a parent directory of your choice

It's better not to put the bindings in Red's directories at all. 
They are isolated from each other and should be handled that way

But in the c-libraray.reds file that is in the zip file with the 
examples I downloaded from Fosil there is this: #include %../common/common.reds 
- but there is no common folder or common.reds folder in that download 
from Fosil - so where does it find this common.reds file that it 
is trying to access if it is not the one that is part of Red/System?

Yes, that's why I asked you if you had installed the extra Red-common 
repository, from the Announce channel

Ah! Sorry Kaj - I missed that vital bit of info - I will go and find 
it  now

Welcome to dependency hell ;-)