If I understand your goal, which seems a little bit hidden by your code
example including extraneous stuff, the code below may do what you want.
The input to show_some should be an array which lists the indices by child
number of what to show.  If instead you meant that individual parts of
labels, like this part is called "widget" and that part is called "hinge"
then to do that, the individual parts need to respond to a signal to
display themselves or not.  You can send information into the child about
its context that it can use to decide if it should display, but you cannot
get information out of the child.

module show_some(show){
for(i=[0:1:$children-1])
if(search([i],show)!=[[]]){
echo(i);
children(i);
}
}

show_some([0,2]){
text("0");
right(10)text("1");
right(20)text("2");
}

On Fri, May 9, 2025 at 5:04 AM Roel Vanhout via Discuss <
discuss@lists.openscad.org> wrote:

On 5/9/2025 2:04 AM, Roel Vanhout via Discuss wrote:

That would be me.

OK.  What I've usually done for that case is to create a new "part" that
includes multiple subcomponents, e.g.

item("part12") { part1(); part2(); }

But what you're asking for is plausible and more flexible.

There's no way to get information out of a module.  (More generally,
there's no way to get information out of a scope, except function return
values.)

If you look carefully, you'll see that item() is the thing that decides
whether or not to build a particular item.  Every item() gets called,
but only the selected one actually generates anything.  Here's the key part:

You'd need to change that "if" to add a test for whether "name" is in
the $content_selected array.  You could use search() as Adrian suggests,
or something like

function inlist(list, val) = len([for (v = list) if (v == val) 1 ]) > 0;


... || (is_list($content_selected) && inlist($content_selected, name))

If you do this, if you look at contents() you will see that it will
generate all of the selected items at the origin.  It seems like that's
probably what you want, but if not then you'll need more work there.

On 5/9/25 15:12, Jordan Brown via Discuss wrote:

I have done some of that, by making a module that includes the modules I
want, building by variable passed, data which builds a 5 part hotend
carriage for a 3d printer, or breaks it down into 5 individual pieces
that print separately in one printing, then assemble. The method seems
quite amenable to making mulltiple parts whose bolt holes align
perfectly.  And I do that with the weekly AppImage build, no libraries
called in. Pure OpenSCAD.

That is controlled by the scope of the variable as its defined.

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis

Libraries are pure OpenSCAD, just organized so that many people can take
advantage of the code

On 5/9/2025 9:41 PM, gene heskett via Discuss wrote:

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 5/10/25 07:10, Jon Bondy via Discuss wrote:

However Jon, libraries such as BOSL2, don't have the on screen quick
reference to tell one how to use them w/o resorting to a web browser on
another workspace. This messes with my "train of thought" and usually
results in my doing what I want w/o the help of the library. Blame some
of that on my 90 yo thinker & you will be pretty close to the mark. That
and 80 years of chasing electrons to make them do what I want.  That I
have been rather good at.  Besides, gotta have something to keep me out
of the bars in my dotage since my diabetes precludes that.  ;o)>

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis

Gene:

Understood, but were you to create a library to perform the magic you
described, then the rest of us would benefit.

I find most libraries to not be worth the time to learn.  BOSL2 has such
excellent documentation, with useful examples, that it is the first
place I go when I need something.  For example, I just discovered their
hinges, with gratitude.

Jon

On 5/10/2025 8:38 AM, gene heskett via Discuss wrote:

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 5/10/25 09:40, Jon Bondy wrote:

Likely true. but the creation is so easy.  The hard part is the first
bolt, then another module creates 4 of them, usually in a rectangular
spacing pattern. Then when you need that bolt pattern, include the 4
bolt version, rotating and translating it to where ever its needed. 10
minutes to put that bolt pattern anyplace its needed once written. 
Working in PETG, I don't consider the threads. The printer makes a hole
.45mm smaller than specified, so I increase the size of the hole by .5mm
for clearance, or .2mm if its to be self threading,  usually for 3 or 4
mm cap screws. At 5 threads into the plastic its self locking and well
threaded. I could likely pull the cap off the screw before the threads
it makes in the PETG fail.

I didn't know it could do that. When I needed to make a 7 degree to do a
buttress thread with the side of a .0625" by .250" round nose mill,
using the rounded nose as a fillet, I considered a hinge under the
spindle motor, something I could adjust, but given the inevitable slop,
I wound up just making a solid 7 degree wedge, which 3 years later is
still installed. Makes beautiful buttress threads in hard maple using
gcode I wrote myself.  I write ALL the gcode I need, no CAD involved.  I
do use a bit of python to cut any sized holes on the milling machine
which is as close to CAD as I get. linuxcnc is very picky about arcs, it
its off a micron it won't accept it.

That box we're supposed to stay in?  It got cut up and fed to the
fireplace back in the late 40's of the last century. . .  OpenSCAD fits
my thinking methods, flawlessly, thank you ALL.

The only thing it doesn't do is Einsteins relativity, but our machines
don't move fast enough to matter. But I had to explain that to the FCC
in regard to a time distortion we didn't know how to precompensate for
back in the 1970's, imposed by the klystron amplifiers us UHF
broadcasters used at the time.  We've since developed both shorter
amplifiers and precomp methods, which work or we'd not have digital hdtv
today.  Adequately explaining that would take a couple more paragraphs,
but has to do with an electron beam about 4 feet long and the mass
increase of said electron when moving at a measurable small fraction of
C speed due to the 20 kilovolts applied. The tube length effectively
grows with the power level of the then NTSC signal.. All of which is
off-topic for this list.  But I can continue in more detail if there is
an interest.

[...]

Cheers, Gene Heskett, CET.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis

Ah thank you, sorry I had forgotten that it came from you, but I have
used it many times since, very useful, thanks :) I did post the entire
thing to give some context, it seems to me that something like this
would be so universally required that it may warrant submission to
BOSL? It feels very closely related to distributors.scad from there.

Yes indeed, I had gotten this far, but the devil is in the details of
the 'more work' mentioned :) I do want to layout separate pieces in a
grid just like if these modules would be the only ones that are listed
(so if I have 4 items, and I want to show the 3rd and 4th, they would
be laid out on the X axis, instead of being offset along the Y axis
and the spots where items 1 and 2 would be, left blank), but if I add
only the appropriate is_list() and inlist() checks to both the
contents() and the item() modules, the parts that are not selected are
just skipped over and their spot in the grid left blank, so to speak.
(My next enhancement would be to add multiple layout modes - in a
line, in a grid, maybe some convenient way to do exploded views of
assemblies). But it seems like a fundamental problem that there is no
way to get the 'name' from an item() inside contents(), and there is
also no way to keep track of the howmanyeth item we're are currently
at within contents() (so that we can pass it to item() with a $
variable), which are overall the two main approaches I think exist.
I've played around with various workarounds with wrapper modules and
so on but they all require passing extra information to either
contents() or item() at some point, which severely degrades usability
because you have to repeat your selection of items.

My next usability thing would be to only have to specify module names
ones, so instead of having to write item("hook") hook(); for each
part, I could just write item("hook") and the item() module would
instantiate a module named "hook", but again there doesn't seem to be
a way to do this sort of dynamic instantiation of modules. I think it
(as well as my previous, original problem) could be solved with some
metaprogramming in a preprocessor, but I guess it would be too much to
ask to have one of those included in OpenSCAD :)

regards

Roel

On 5/14/2025 1:27 AM, Roel Vanhout wrote:

Yeah, can't avoid that with this scheme.  Only the child item() subtrees
know whether or not they are generated, and only the parent contents()
knows where the children are positioned.

The grid supported by contents() is rudimentary, a way to provide an
overview of the parts.

Exploded views are not as easy as you might think, because the position
of everything depends on the position of everything else, sometimes in
contradictory ways.

Correct, there is no way.  OpenSCAD children (and in general OpenSCAD
scopes) are black holes; no information can escape.

Also correct, again because the children are black holes and won't tell
you whether or not they are being generated.

You can imagine giving the list of names to contents() and having it do
the on/off selection.  But that's ugly from a modularity standpoint,
having to maintain two parallel lists.  (As you mention elsewhere.)

PR#4478 (don't hold your breath) offers some additional options - in
particular, using module references.

There is not.