Hi Thomas,
You wrote:
>Does Kaffe support JIT on PPC ?
Yes, it does. For Linux, Darwin, MachTen, etc.
It needs a full "config" development for OS4, followed by a port of the JIT
for OS4. Like what I did for the M68K version and OS3.9.
cheers,
Well, I've now some free time (unless They find something else for me to
do), so I can do some more work on OOo at last.
What has highest priority? Is there anything specific, or should I have a
go at getting Kaffe working under OS4?
[I have this dream of getting Kaffe to create a cache of already-jitted
methods, to avoid the translation time whenever a method is reloaded. It
could go well with Amiga shared libraries, as long as I can avoid the GPL
virus.]
cheers
It's time to get started on some more work. I mentioned yesterday that I
had an idea for something that might bring in a couple of more
programmers. Apoligies in advance for stepping on anyones toes here, but
I think we've been a bit too silent lately. And that isn't good in the
long run...
The idea is this:
Construct a proper dependency graph for OO.org, and then use this to
construct an "enabling end-products" graph. I'm not sure what to call
it, but I basically mean taking the dependecies we have in OO.org, and
figuring out what other useful and/or cool pieces of software will be
that much closer to a port if/when each dependency is taken care of.
That way both we and people who come across some kind of source they
want to port will know where to start.
I'm thinking of doing it as a graphical diagram as well as a relational
database (because I do both of those things a lot already). That way you
could search in the database for a piece of software to see what
dependencies there are and if anyone is working on them, or search for a
dependency and see how useful it is.
Does it sound useful? I think it would make it a lot easier for people
like Captain Moo Moo and Joppe, who were both looking for projects to
feed to students. Just putting all of them into the OO.org pool would be
too much noise and not enough gain, but splitting them into teams to
tackle smaller (but still important) parts of the project or even build
other projects on top of our dependencies to extend the software base of
the Amiga would be great.
So we need to concentrate more on those dependencies. I know Henning has
spent a shitload of time on this already, I would actually like to know
at this point who _is a programmer_ and _has some time this coming
weekend_ (and if yes on both points, _when_). Yes, an IRC schedule. We
won't be needing non-programmers, but an hour or two of nitty gritty
code reviews and notes taken which can be built into a database would be
a good place to start if we want to attract some more attention. For
one, it would serve as a kind of "progress report" to the outside world.
I'll be sketching down the outline of a database tomorrow, if any
programmers has some free time between friday evening and monday
morning, european time, send me a PRIVATE EMAIL with the fixed
world-clock time ( http://www.timeanddate.com/worldclock/fixedform.html )
like this:
start: url-specifying-start-of-period
stop: url-specifying-end-of-period
repeat for each period
If you have better ideas for scheduling practices, please make note of
them IN ADDITION to following my wish. At least this once, we'll see if
we use your ideas in the future, ok? :-)
And I repeat, we really only need people who can go over code/search
elsewhere to look for obvious dependencies. I'm sure we won't find all
dependecies (and their order/structure, which is just as important) just
yet, but maybe we can start filling in something.
It would be awesome if someone could find a member of an existing oo.org
porting project team and get some input there. That would actually help
a lot.
Of course, if people choose to ignore my idea then that is perfectly
acceptable, but I hope I'll then at least get a response indicating that
the email has been read but the idea rejected. Even though it ended up
as a rather long email in the end...
--
We'll double-cross that bridge when we get there.
