Rinat Sadretdinow wrote in a message to All:

 RS> Does anybody know anything about undocumented statements of
 RS> LINK386 and ILINK? There is a list of all reserved words in
 RS> the Visual Age's linker guide and it contains a number of
 RS> words not described anywhere. Below I include a cut from the
 RS> guide where I've marked all such words with an asterisk: 
* * *
 RS> Some of these words are self-explaining, for example
 RS> [NO]EXPANDDOWN, LONGNAMES, etc. But another are not. Any
 RS> suggestions?

Most of these keywords are useful only with regard to segmented Ring 0
code, such as device drivers, which are technically not supported under
LINK386.  Some of the keywords simply set attribute bits in the EXE header;
see the EXE386.H file in the OS/2 Toolkit for more information,
specifically the O32_FLAGS definition.  Some of the keywords are simply
shortcuts for setting more than one primitive flag bit, such as CONTIGUOUS.

Some keywords are supported only for backward compatibility.  For example,
LONGNAMES is pointless because LINK386 only makes 32-bit targets, and all
32-bit targets will see long file names.  The keyword, however, is a
vestigial carryover from the 16-bit linker, needed in order to protect OS/2
1.0 and 1.1 applications from seeing the long file names introduced with
OS/2 1.2.

IBM also intended the LX-format module to be used across platforms,
including the PowerPC.  This is why such things as endian-ness codes are
provided.
 
-- Mike


---