libgl.a questions

[Kipp Hickman]

In article <9104231603.AA15231 at noname>, james at adt.UUCP (james cerrato) writes:
|> We ship a graphics product to clients using many different SGI machines,
|> running diffent releases of IRIX.  In order to deliver the product without
|> first having built the executable on each configuration [Hardware/OS] first,
|> we just ship a library of our object files.  Then our installation script
|> links our library with -lgl at the client's machine.  Will this strategy
|> work even on machines that do not have installed the development option?
|> 
|> Also if anyone can give me information on how to use shared libraries,
|> and what the benefits are, or a pointer to where to find information about
|> them in the manuals it would be greatly appreciated.
|> 
|> Thanks for your time.
|> 
|> James Cerrato
|> Associative Design Technology
|> Westboro MA

Ahem.  I don't know the answer to the first part - if systems without the development option will be able to construct your application...However, you can construct it yourself by linking it with the shared gl library.  Speaking of that, there will no longer be an unshared library in future releases.  Its a bad evil nasty icky broken thing that should never have been shipped because it leads to problems beyond belief.  In any case, if you link your program with the shared gl it will run on ALL 4D platforms







, thus (I am guessing) solving two problems with one solution.

So here is how you use shared libraries:

	$(CC) -o whatever ... -lgl_s -lm -lc_s

and away you go.

					kipp hickman
					silicon graphics inc.
-----
My opinions are SGI's opinions, since everybody else is weasling out :-)