Re: [MacPerl] Problem with tie and DB_FILE

[Matthias Ulrich Neeracher <neeri@iis.ee.ethz.ch>]

oszarazi@cuug.ab.ca (???) writes:
>On Mon, 16 Dec 1996, Dr. M. Hartje wrote:
>> you want to use MacPerl 5 on a Mac Plus?
>> 
>> If you have any experience to bring it to work, please report on it.
>
>Nope.  It's not working.  

Yes. Sorry.

>> As Matthias Neeri answered to my question a few month ago, the port
>> of macPerl 5.x need a 68020 at minimum hardware configuration.
>
>Ahh.  A 68020.  Isn't it possible to mimic a 68020 with some form of
>extension to the ROM routines or something?

Not reasonably in this case. Absence of an MMU is much easier to deal with.

>I can't understand how a chip would impact the running of an application on a
>Macintosh.

OK, here's the deal:

MacPerl has to deal with massive amounts of code and data. The 68000's
addressing formats are limited to 16 bit offsets in instructions. To use 32 bit
offsets, as is necessary in MacPerl, 68000 processors have to generate rather
unpleasant code workarounds.

68020 and later processors have additional addressing modes allowing 32 bit
offsets. Thus, turning on 68020 mode in the compiler generates smaller and
faster code for MacPerl. I thought at the time I introduced MacPerl5 that not
too many 68000 machines would be in operation, which turned out to be a not
entirely correct prediction.

Building an 68000 version would be simple, but I don't have time to maintain
yet another distribution. Sorry.

>> So I have ggod experience using The Perl 4.13 and 4.18 on a Mac SE
>> with Sys 7.1 and 4 MByte Ram. It is sufficient for most of the
>> problems I wnat to solve on this machine.
>
>Right now, I'm running 4.13.  MacPerl 4.18 won't run for me either!!
>(Plus ... 4Mb RAM ... Sys 7.0+).

Hmm, 4.1.8 might run with SoftFPU or so.

Matthias

-----
Matthias Neeracher                                   neeri@iis.ethz.ch
   "These days, though, you have to be pretty technical before you can 
    even aspire to crudeness." -- William Gibson, _Johnny Mnemonic_