home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!news.hawaii.edu!s.u-tokyo!ccut!wnoc-tyo-news!nec-tyo!nec-gw!sgiblab!sdd.hp.com!decwrl!concert!ais.com!bruce
- From: bruce@ais.com (Bruce C. Wright)
- Newsgroups: vmsnet.internals
- Subject: Re: Dream extensions for MACRO-32
- Message-ID: <1992Nov4.145733.5817@ais.com>
- Date: 4 Nov 92 14:57:33 GMT
- References: <00962A13.84566C80.28833@WKUVX1.BITNET> <1992Nov3.211427.21777@nntpd.lkg.dec.com>
- Organization: Applied Information Systems, Chapel Hill, NC
- Lines: 44
-
- In article <1992Nov3.211427.21777@nntpd.lkg.dec.com>, critz@star.enet.dec.com (Richard Critz) writes:
- >
- > In article <1992Nov3.093855.5816@ais.com>, bruce@ais.com (Bruce C. Wright) writes:
- > |>I would be very skeptical about changes that, for example, hid the number
- > |>of registers from the programmer, and inserted extra instructions to save
- > |>and restore registers, etc -- I can forsee too many possible problems in
- > |>driver-level code where you often have to use specific instruction types
- > |>when you refer to device space, or where timing can be very important.
- > [...]
- >
- > This has effectively already happened. MACRO32 is a compiled language on
- > OpenVMS AXP(tm)* (known in the real world as Alpha). The compiler does all
- > sorts of things behind your back including the allocation of scratch
- > registers. Some assumptions about stack layout and about where and how
- > one can legally exit a routine which are valid on the VAX are not valid
- > on the new architecture.
-
- Yes, but we were talking about versions of MACRO32 for the VAX
- architecture too. And on that architecture there are some situations
- where you have to be very careful about what you can do in device
- space, particularly.
-
- The same goes for the byte/word displacements, and for the I^ and S^
- literals.
-
- > |>But why bring this up now, when the VAX is almost history? Some of us
- > |>have wanted this type of thing for over 10 years, but DEC has never done
- > |>anything about it. I suspect that the problem is the typical DEC attitude
- > |>of NIH (= Not Invented Here) -- a certain other competing operating system
- > |>for the VAX has had this since forever.
- >
- > If you read the original message again, you'll see that he brought it up
- > in response to the fact that MACRO32 is a compiled language on OpenVMS AXP.
- > To the original poster, it's still an assembled language on OpenVMS VAX and
- > the assembler is in maintenance mode and has been for years.
-
- This is a perfect example of exactly what I was saying -- DEC is not
- interested in listening to what customers want, nor in what anyone
- outside of DEC has done.
-
- It pains me to say it since I've been working with DEC gear for over
- 15 years, but with an attitude like that you deserve to lose.
-
- Bruce C. Wright
-