Wed, 13 Aug 1997 18:30:56 +0200 (MET DST)
I've promised to speed up the release rate, and keep the list more chatty,
so here is release 0.0.0.34 and lots of topics...
The main novelty in ...34 is that the TUNESADM script now works,
and allows easy management of three versions of TUNES (plus backups):
* a reference "official release"
* a local development version
* an external development version to merge with
This allows developers and maintainers to exchange and manage patches
wrt (with respect to) previous official TUNES release (or older)
while having their own development version.
I haven't experimented with diff3 yet, since I have no contribution
to merge with yet.
Another interesting feature of TUNESADM is that it
manages execution of remote commands (zsh scripts) through PGP-signed mails,
so I can run commands on the machines at ENS without having to be logged in
at interurban rates (from holiday site). Of course, the first command I
implemented was full automatic update of the TUNES' server site
from a compressed/encoded mail containing just the official patch.
I admit I already had to modify TUNESADM since the release,
though it ran well on my Linux homebox, because the ENS machines
run under ScumOS with its inferior command-line tools
(as compared to Linux/GNU tools, or anything -- down with software hoarding!).
Of course, you'll say, this mostly concerns active developers,
subproject maintainers, and administrators of Tunes WWW/FTP mirrors,
which currently means only me. However, if any of you is interested,
I may add him to the list of people receiving automatic update
of TUNES at home via e-mail. The first testers will probably have to
be bug-reporters, too ;-) but it already works "well enough for me"
Note that the code could well be used for other purposes than TUNES,
if anyone feels like packaging it. I wouldn't do that myself,
since there are too many dependencies and options to manage
(zsh, GNU *utils, sed, perl, procmail, pgp, tar, gzip, bzip, unzip, etc).
The only code I feel like releasing really is Tunes.
I've also cleaned-up somewhat the Makefile's, autoconf.sh, depend.pl,
and such, so that though not much actual running code is here,
the infrastructure to develop it is now ready.
Let's now proceed with actual coding.
I've looked at Flux OS-kit 0.60 (thanks again, Nemo), and it looks like
the trick about it would be to develop OTOP then port OTOP to it and gain
impressive performance gains by having direct MMU access
(including dirty bits) instead of inefficient mmap() and mprotect()
or software memory management. So let's focus on OTOP for now...
I'm still looking for the right package to manipulate C code
from Scheme. Unhappily, I've left ~twaddle home, and don't feel like
d/l'ing it over interurban 14.4kbps. I'll have to go through sources
for various Scheme->C compilers, too, but the ones I've seen seem
to output C code through format w/o actually manipulating it like Ctool does.
As for the semantics of the HLL, well, I'll use the experience
I'm acquiring writing a few interpreters for my DEA...
As usual, there's been maintenance in the docs and WWW pages:
enhanced READMEs, new stuff in the Review/ subproject, etc.
However, it becomes everyday more blatant that HTML is much too low-level
for manual management, and we need a way to automatically generate it.
I already began to use the SGML Tools from the Linux HOWTOs
(with my experience as Assembly-HOWTO maintainer),
but even though they are better than nothing, it's not great.
The most interesting track seems to be DSSSL, but I've not found
the docs for it yet, and again, I'm too far away to take a look before some
time. Suggestions, pointers, etc, welcome.
I still received no feedback (either positive or negative) about
my idea of a Tunes party in Paris in late september. Are all french
Tunesers not reading, being in holiday? [other Tunesers sure welcome,
if can be present].
Finally, I won't tour to the States this year, by lack of time,
finance, preparation, but will go next summer, so that's one year to
plan destinations et al...
PS: have you seen the ACL2 theorem prover? Impressive -- has been used
to prove correctness of part of all of actual CPU designs,
as well as theorems in higher mathematics...
n-uf sed. Have a nice summer, and good hacking.
== Faré -=- (FR) François-René Rideau -=- (VN) Ð£ng-Vû Bân -=- firstname.lastname@example.org ==
Join a project for a free reflective computing system! | 6 rue Augustin Thierry
TUNES is a Useful, Not Expedient System. | 75019 PARIS FRANCE
http://www.eleves.ens.fr:8080/home/rideau/Tunes/ -=- Reflection&Cybernethics ==