CM> Craig Swanson doth sayeth in a message unto All:

CS> Frank Andress called me today to state that the behavior I CS>
have observed regarding Warp 2 no longer accepting long file CS> name
DLL's is not a bug but is in fact a fix to a bug.

 CM>   Excuse me?? A bug fix? I don't think so..

I'd agree that this change was one way to fix the bug of DLL's with the
first 8 identical characters of a long file name being regarded as the same
DLL.  However, I don't really think this bug fix was in the spirit of OS/2
-- i.e., it added a limit rather than removed one.

CS> would be problems.  The fix results in names longer than 8 CS>
characters (plus the DLL suffix) no longer being accepted.

 CM>   Sounds like a quick hack to patch a problem more than a real bug fix..

While I'm the one of the last ones to defend IBM when it comes to fixing
bugs in OS/2 (thanks to their bad track record on fixing bugs I've
reported), I don't see that restricting DLL's to 8.3 names is a major
problem.  However, this restriction needs to be more clearly documented
someplace and IBM's tools should enforce the restriction.

In the long run, however, there needs to be a way to have multiple DLL's
with the same name loaded.  What happens when vendor A requires version
1.01 of a DLL and vendor B requires version 1.02 of the same DLL?  As it
stands right now, situations like this essentially mean that OS/2 can only
reliably run one of the two programs at any one time.


--- Maximus/2 2.01wb