[Info-vax] Native compilers

[VAXman- at SendSpamHere.ORG]

In article <sv7on8$ljq$1 at news.misty.com>, Johnny Billquist <bqt at softjar.se> writes:
>On 2022-02-23 21:02, VAXman- at SendSpamHere.ORG wrote:
>> In article <sv5ias$ute$1 at news.misty.com>, Johnny Billquist <bqt at softjar.se> writes:
>>> On 2022-02-22 16:06, VAXman- at SendSpamHere.ORG wrote:
>>>> There is no .INCLUDE directive in Macro-32!  It is not referenced in the
>>>> VAX Macro LRM either.  Adding a ".INCLUDE /whatever/" elicits the error:
>>>>
>>>> 	.INCLUDE	/whatever/
>>>> ^
>>>> %AMAC-E-UNRECSTMT, unrecognized statement
>>>> at line number 173 in file SYS$SYSROOT:[SYSMGR]X.MAR;3
>>>>
>>>>
>>>>> Or at least this is what I *know* is true in Macro-11, and while it's
>>>>> been quite a while since I played with Macro-32, I'm fairly certain it's
>>>>> the same there.
>>>>
>>>> 'Tis not!
>>>
>>> Ok. This time I'm really surprised. Makes one wonder why this was
>>> removed in Macro-32, which certainly otherwise feels like a spiritual
>>> successor to Macro-11.
>>>
>>> And I certainly believed I had used it in the dim past when I played
>>> more on VAXen. Brains develop bit rot...
>> 
>> Personally, I prefer the macro library mechanism.  If you need to ".INCLUDE"
>> put it in a macro library, reference the library (.LIBRARY) and then callout
>> the module.  The idea is reusable macros.  One time Macro-32 should simply be
>> put into the source.
>
>They have different uses, and are not interchangeable.
>I certainly like the macro libraries, but they cannot be used for things 
>like some prefix file you want to add to the compilation. Which is why 
>..include is also good. Of course, you can also do this on the command 
>line, which is what I believe John Reagan was hinting at.
>
>But I find it a bit annoying to have to remember to throw those things 
>in on the command line, so when possible, I use .include instead.
>
>Typical things are stuff like you have some configuration step that 
>based on your choices creates a prefix file with various symbols 
>defined. Which is then included in with the compilation, and the source 
>have various conditional parts that are included or not (or modified) 
>based on what's in the prefix file.

VMS provides a file for multiple target (VAX, Alpha, Itanium and I would
assume for the X86 too) that defines various conditional symbols for the
different architectures.

In my DESCRIP.MMS file, I create a library (SYS$LIBRARY:ARCH.MLB) and I
then sandwich the contents of SYS$LIBRARY:ARCH_DEFS.MAR inside of .MACRO
and .ENDM directive and insert that in the ARCH.MLB.  Then, my code can
incorporate this with:

	.LIBRARY	"SYS$LIBRARY:ARCH.MLB"
	.ARCH_DEFS



>You could, of course, create a macro out of the whole thing, and put 
>that in a library, and then the source refers to that created library, 
>invokes the known macro that defines all the choices made, and then you 
>are good. But that seems like a very complicated way of doing things here...

MMS makes this virtually effortless, IMNSHO.  If you can edit your source
to add a ".INCLUDE", you can edit it to callout a macro.



>(There is so much conditional code in RSX it's almost ridiculous. 
>Without such prefix files selecting which way to build things, it would 
>become horrible. Hello LB:[11,10]RSXMC.MAC)

comp.os.vms
--------^^^


-- 
VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)ORG

I speak to machines with the voice of humanity.