What would extrude_for() do?

On Wed, Sep 27, 2023 at 7:08 AM Martin Axelsen kharar@gmail.com wrote:

On Wed, Sep 27, 2023, 7:08 PM Adrian Mariano avm4@cornell.edu wrote:

What I want to say is that we can make hull + union as fast as hull without
union, so we should focus on semantics but not performance.

The optimization is not implemented for now (in manifold) because at the
time when I implemented tree rewriting support for manifold, we did not
have hull, but I can do it if you want.

In fact there is a tracking issue for manifold about improving csg_tree
performance, but it requires refactoring the csg_tree and I am currently
working on other stuff (fast 3D offsetting). csg_tree optimization is of
low priority now because users seem to be pretty satisfied with the
performance right now, so I would rather focus on implementing other things
which can provide larger gains. 3D offsetting seems to be the main usecase
for minkowski, and minkowski is terribly slow...

On 9/27/2023 2:57 AM, nop head wrote:

Maybe I just don't understand, but I don't think it does.  (And it
really can't, without huge changes in semantics.)

With 2023.09.25, with lazy-union on, manifold on or off:

module c() {
    echo($children);
    children(0);
    children(1);
}

c() {
    for(i=[1:9]) cube(i);
}

yields

ECHO: 1
WARNING: Children index (1) out of bounds (1 children) in file , line 4

... which it kind of has to, because child modules aren't executed until
you call children().  For for() and other modules to be able to return
individually accessible children would require that the child modules
all be executed before the parent is executed.

On Wed, Sep 27, 2023, 11:45 PM Jordan Brown openscad@jordan.maileater.net
wrote:

I wonder what semantics changes we need to make it work? I think openscad
is purely functional with no side effects (except echo?), so different
evaluation order should not affect the result?

You can pass $ variables to the child so you could change the number of
iterations of the for loop which would be circular if you based it on
$children!

On Wed, 27 Sept 2023 at 17:14, pca006132 john.lck40@gmail.com wrote:

On Thu, Sep 28, 2023, 12:27 AM nop head nop.head@gmail.com wrote:

Thanks! I guess I should read more about the semantics of OpenSCAD before
discussing such semantics issue... This is too non-standard that my
experience in other languages does not help.

Maybe this will make me sound like a jerk, but from a programming language
point of view I think it would be much better to create something like a
OpenSCAD2, put the old behavior in a compatibility mode, and redesign the
language (if users still think it is desirable to have a domain specific
language for 3D modelling, instead of using general purpose languages). It
seems to me that there are so many design issues and limitations that we
cannot fix without breaking some existing code.

There is a project afoot to make a Python version.

On Wed, Sep 27, 2023 at 11:42 AM pca006132 john.lck40@gmail.com wrote:

Jordan, my understanding, which I haven't tried to verify, is that the lazy
union option isn't fully implemented, and in particular, that it doesn't
work for passing children into user-defined modules.  That's why your test
fails.  But if it were fully implemented then your test would work and it
could simplify the xor() call.

On Wed, Sep 27, 2023 at 11:46 AM Jordan Brown openscad@jordan.maileater.net
wrote:

I’m still not up to speed apparently. lazy-union is a feature that can be enabled or disabled? What precisely is the semantics of this feature? And the semantics of its absence?

-Bob
Tucson AZ

On Sep 27, 2023, at 08:45, Jordan Brown openscad@jordan.maileater.net wrote:

On 9/27/2023 2:57 AM, nop head wrote:

Maybe I just don't understand, but I don't think it does.  (And it really can't, without huge changes in semantics.)

With 2023.09.25, with lazy-union on, manifold on or off:
module c() {
echo($children);
children(0);
children(1);
}

c() {
for(i=[1:9]) cube(i);
}
yields

ECHO: 1
WARNING: Children index (1) out of bounds (1 children) in file , line 4

... which it kind of has to, because child modules aren't executed until you call children().  For for() and other modules to be able to return individually accessible children would require that the child modules all be executed before the parent is executed.

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Lazy union is a feature you can enable in the nightly build dev versions.
It's actually been available for a year or two.

As noted earlier, it should be called no-implicit-union.  When you write a
for() you get an implicit union of the objects generated in each iteration
of the loop.  So code like:

parent() for(i=[0:9]) {mything(i)};

will pass a single object to parent, the union of ten copies of mything.
The parent was hoping to see ten children and do something different with
them you're out of luck.

If we get rid of implicit unions by turning "lazy union" on then the same
code should pass ten separate instances of mything to the parent.  The
parent will have ten children.  It can union them if it wants to, but could
also do something else.

My understanding, as noted earlier, is that "lazy union" isn't actually
fully implemented, so even with it on, you still get implicit unions for
children passed to user-written modules.  But in the case of hull you can
avoid the union, which, before the "manifold" option was available, meant
your code could run 100x faster.  Now it's just 6x faster.

One thing that does happen with the current implementation is that at the
top level there is an implicit union normally so the output STL file
contains one, single object.  If you enable lazy union then the output can
contain multiple, possibly intersecting, objects.

On Wed, Sep 27, 2023 at 5:08 PM Bob Carlson bob@rjcarlson.com wrote: