92 lines
3.5 KiB
Plaintext
92 lines
3.5 KiB
Plaintext
From: morse at harborcom.net (Kevin Dahlhausen)
|
|
Date: Fri, 09 Apr 1999 11:37:54 GMT
|
|
Subject: Python for embedded controllers?
|
|
References: <F9HuLr.KwL@world.std.com>
|
|
Message-ID: <370de606.77891472@news.oh.verio.com>
|
|
Content-Length: 3239
|
|
X-UID: 1415
|
|
|
|
I know you already mentioned the mentioning of Python on this mailing list, but
|
|
this was really exciting to me and rest of the snake pit might like it too:
|
|
|
|
|
|
SO, maybe when it's available, their microPython source can serve
|
|
as the basis for other embedded systems.
|
|
|
|
---------------------------------------------------
|
|
|
|
Date sent: Thu, 8 Apr 1999 11:34:15 -0700
|
|
To: ucsimm at uClinux.org
|
|
From: Michael Gorlick <gorlick at aero.org>
|
|
Subject: Re: [uCsimm] Scripting Languages on uCSIMM
|
|
Send reply to: ucsimm at uClinux.org
|
|
|
|
At 10:03 AM -0700 4/8/99, William Ware wrote:
|
|
|
|
>Python uses a compiled bytecode which is executed by a virtual
|
|
>machine, like Java. It might be nice to be able to bring up just
|
|
>the VM on a target board, and download compiled bytecode to it,
|
|
>rather than running a full interpreter on the board. I would
|
|
>also think that for a cpu with no floating-point unit (I don't
|
|
>think the dragonball has one), you'd want a Python build that
|
|
>avoids any use of floats or doubles. I haven't looked closely
|
|
>enough at the Python source to see how feasible that would be.
|
|
>The few times I've tried building a 68K gcc cross-compiler, I've
|
|
>never been able to get the floating-point library (libgcc1.s?)
|
|
>to plug in correctly. I don
|
|
>t have any experience trying to build ecgs as a cross-compiler.
|
|
|
|
Be assured that we are acutely aware of the limitations of the DragonBall
|
|
processor and have been completely ruthless in stripping it down to the
|
|
bare minimum
|
|
(truth in advertising requires that I disclose that in the interest of time
|
|
a few extraneous bits and pieces have been left in --- these were components
|
|
that were easier to ignore than remove). We estimate that with more work
|
|
we can reduce
|
|
our version of Python down from its current measly 124 kB to about 110 kB.
|
|
Further
|
|
reductions may be possible but with correspondingly greater effort.
|
|
|
|
I have also been a little sloppy with terminology. We are not running the
|
|
full interpreter
|
|
but just the Python virtual machine and a spare framework around it. For
|
|
example,
|
|
the interactive, command-line front end is history along with the builtin
|
|
compiler.
|
|
Users will use a "full up" Python interpreter to produce VM bytecode files
|
|
(*.pyc)
|
|
which will then be downloaded onto the Palm Pilot.
|
|
|
|
On another note, many interested users have contacted me personally to
|
|
enquire further or
|
|
to offer their assistance. Your interest and enthusiasm is greatly
|
|
appreciated, however,
|
|
at this point in time no further information is available. We are doing
|
|
this as quickly as our
|
|
congested and conflicted schedules permit. We anticipate having an alpha
|
|
version in house by
|
|
the end of April. We plan to make the source code and prebuilt binaries
|
|
publically available,
|
|
although the exact timing of the distribution is unclear. We will make
|
|
announcements
|
|
on this and other mailing lists when the distribution is packaged and ready
|
|
to go.
|
|
|
|
We recognize the value and potential of Python on platforms like the uCsimm
|
|
and are anxious
|
|
to have it in the hands of users such as yourselves. Please be patient
|
|
with us.
|
|
|
|
__
|
|
_/mg\__
|
|
... /o-----o>
|
|
|
|
|
|
This message resent by the ucsimm at uclinux.org list server
|
|
http://ryeham.ee.ryerson.ca/uClinux
|
|
|
|
|
|
|
|
|
|
|