RTL.inc

[Peter]

Deleted.....

[Charles Pegge]

Hi Peter,

Yes I could reproduce the problem. 'A' was already defined as a global variable in RTL32. I have now removed it. (You can workaround this problem by putting a dot before it '.A:').

I have also made a slight modification to the Assemble so you can use dotted labels as well as colon labels on the same line as an Assembly code instruction.

Code: OxygenBasic

1.  
2.   'USING LABLES ON THE SAME LINE AS INSTRUCTIONS
3.  
4.   .a  mov eax,eax
5.   a:  mov eax,eax
6.   .a: mov eax,eax
7.   .a  print "ok"
8.  

The Oxygen 'in progress' version has been updated.

Charles

[Charles Pegge]

Yes I see what is happening.

With the new Oxygen, you can use 'a:' or '.a' or '.a  :  ' but not '.a:'

I can fix the latter problem though I would not recommend that way of expressing a same-line label.

Code: OxygenBasic

1. sys c
2.  
3.    mov ecx,10
4.    mov eax,0
5. a: inc eax
6.    dec ecx
7.    jg a
8.    mov c,eax
9.  
10.    print c
11.  
12.  

Charles

[Charles Pegge]

I have just posted another Oxygen 'in-progress' version which also accommodates '.a:'

Code: OxygenBasic

1.  
2. sys c
3.  
4.     mov ecx,10
5.     mov eax,0
6. .a: inc eax
7.     dec ecx
8.     jg a
9.     mov c,eax
10.  
11.     print c
12.  

Charles

[Charles Pegge]

[ebx+488] is a low level call vector to ReadFile. So I think there a name conflict with your software somewhere.

You could try #undef ReadFile after RTL32 and see if that fixes the problem.

Charles

[Charles Pegge]

This is the first time anyone but me has used the RTL.

I think we may have further name conflicts causing the problems you see.

It may be necessary to disable the rule that distinguishes a procedure call with a colon attached, from a label with a colon attached. This is causing the RTL namespace to intrude on your Application namespace with regard to labels.

Charles

[Charles Pegge]

Yes certainly Peter. It is very important that the path for creating DLLs should be as smooth as possible.

The main issue here is that we cannot safely disambiguate between a procedure call and a label when colons are stuck on the end. Therefore the first word in a line, if it has a colon, must be assumed to be a label.

If we stick to this rule then it should eliminate many potential name conflicts when using libraries.

Charles