World: r3wp

rebolers are all deviants.

On a good day :)

equal? a b

diverge? a b

I missed a lot of this conversation, but saw some mention of NAN 
and INF.  Exceptions by default seem like the natural choice.  Anything 
different could be done with refinements.  divide/symbolic 1 0 ... 
or something?

That is what I was saying about ? in REBOL before. EQUAL? is short 
for EQUIVALENT, and DIVERGE? would be short fot DIVERGENCE. It's 
nonsense in English, but English is just the base language for REBOL 
function names.

Problem with `diverge?` is that the proposed function will not return 
a boolean, unlike equal?.

Or INDEX?, LENGTH?, SUFFIX?, ...

It returns the first series at the point of divergence.

Yeah, those always have been ill-conceived names.

exactly what I was pointing out earlier.    index, length suffix 
should have their ?  dropped.

Fork, the Inf and NaN discussion has moved to CC bug#1717.

And use -of instead. But that's another discussion.

It's just that I would not repeat that mistake when introducing new 
names.

Make that point in bug#1719, where Carl can see it. He tends to let 
us play out these discussions on our own.

I notice that the disarm stub function which was useful for running 
code written to work in R2 under R3 is not included in A109. Is this 
by design?

Yes, Peter. It was intentionally removed in A108 (cf. http://www.curecode.org/rebol3/ticket.rsp?id=1508)

A small gotcha for people who want to write scripts that run under 
both R2 and R3.

One of many, I fear. We should probably start an "Incompatible changes" 
page on the wiki and collect them.

I'm still not sure this specific stub should be removed.  it will 
invalidate *many* R2 scripts and its not such a big thing to leave... 
as long as it prints out a warning.

it could actually be part of a "compatibility" mode switch.... something 
like a delayed module which only need to import when running r2 code.

The other way round is an option as well.

We can collect all of those in an "r3-backward" module.

(Ignore that. Reading comprehension failed me :)

yes exactly.  an r2-compatibility module

That is a better plan than leaving these anachronisms in R3 by default 
:)

but we need to make sure we collect them all (I gather you've been 
a pack rat about this ;-)>

There are some things that we won't be able to replicate, like evaluation 
rules.

So it'll be more like an R3/Backward :)

I expect that with the R2/Forward additions to R2 there might be 
a useful middle ground.

is import expected for R2/forward?  even if its just for non dll 
stuff?

Yes, as an addon (not part of R2 by default). I wanted to wait for 
the model to be set first, and now (for a110) it will be.

cool.  that is the biggest issue I have now when trying to run simple 
R3 stuff in R2.

yes... I am using modules  :-)

I assumed so, but it's nice to hear that you are using R3 modules 
:)

(a classic case of.... if you can't convince them... use the stuff 
they put all those hours on... its probably better than yours  ;-)

hours -> weeks, months?

It feels like forever. But it is really cool now.

I really like how you've evolved the system and its doing all of 
what I wanted slim to do minus one or two little things, which I 
can easily graft over modules as they are .

The next step is to publish the test suite in DevBase, then finish 
it, then make it work with loadable module systems.

it
 being the test suite.

Look at bugs 1721, 1722 and 1723. They are all fixed in the new version 
that is done but not built yet.

I like 1722

in 1723... did you really have to rewrite stuff, or just use the 
newest code?

The boot exts loader does its own importing, for efficiency. As such, 
it wasn't much code to rewrite.

I posted the ticket after I did the work.

Just added this note to #1722: "Note: This ticket is not a dup of 
#1177. We made the same error again with a new module system. We 
have a better solution this time though."

it would be nice if the rename options supports none, to allow explicit 
reloading of modules going thru the whole motion.

And it was a two-word fix :)

since you are telling it none, its not a security hole you chose 
to ignore the name.