• veb (11/11) Jan 04 2010 Hello,
• Walter Bright (4/19) Jan 04 2010 Use the lib program to extract the object module from sds.lib that has
• veb (5/5) Jan 04 2010 Thank you, it works very well!
• Walter Bright (3/7) Jan 04 2010 Provide your own definition of the public symbols in it, then they won't...
• veb (9/16) Jan 05 2010 I tryed:

veb <vebveb hotmail.fr> writes:

Hello,

I have to made a program with the small model of dos 16 bits, so it use
SDS.lib to link.

But To use the programme, I need that the 'Program entry point' is before the
segment 0000:8000
Nowaday, my Program entry point is at 0000:8234, so I have a problem

In a detailed map link, I see:
0822EH 08452H 00225H   Module=SDS.lib(c)
So the program begin in Module=SDS.lib(c)

How could I put this module in an other segment (for example 0000:0000 or
0000:6000)?

Walter Bright <newshound1 digitalmars.com> writes:

veb wrote:
 Hello,
 
 I have to made a program with the small model of dos 16 bits, so it use
 SDS.lib to link.
 
 But To use the programme, I need that the 'Program entry point' is before the
 segment 0000:8000
 Nowaday, my Program entry point is at 0000:8234, so I have a problem
 
 In a detailed map link, I see:
 0822EH 08452H 00225H   Module=SDS.lib(c)
 So the program begin in Module=SDS.lib(c)
 
 How could I put this module in an other segment (for example 0000:0000 or
 0000:6000)?

Use the lib program to extract the object module from sds.lib that has 
the entry point, and put that object module first in the command to the 
linker.

veb <vebveb hotmail.fr> writes:

Thank you, it works very well!


An other question: I don't use double number(only float), but:
08B9EH 09B30H 00F93H   Module=SDS.lib(double)
It take some place (3ko, and I'm limit to 64ko code)
How could I compile without the 'double' definition?

Walter Bright <newshound1 digitalmars.com> writes:

veb wrote:
 An other question: I don't use double number(only float), but:
 08B9EH 09B30H 00F93H   Module=SDS.lib(double)
 It take some place (3ko, and I'm limit to 64ko code)
 How could I compile without the 'double' definition?

Provide your own definition of the public symbols in it, then they won't 
get linked in from the library.

veb <vebveb hotmail.fr> writes:

I tryed:
#undef double
#define double float
but it doesn't do anythind, and I tryed to replace every double and long double
in
my  /include to float, but it doesn't change anything.

I think Digital mars is a very good compiler, very fast for compile and to
execute, but the exe are maybe a bit biger and the function gotoxy doesn't
exist.

Thanks for everything
== Extrait de l'article de « Walter Bright (newshound1 digitalmars.com) »
 veb wrote:
 An other question: I don't use double number(only float), but:
 08B9EH 09B30H 00F93H   Module=SDS.lib(double)
 It take some place (3ko, and I'm limit to 64ko code)
 How could I compile without the 'double' definition?

 Provide your own definition of the public symbols in it, then they won't
 get linked in from the library.