On Friday 13 August 2021 08:28:45 Gene Heskett wrote:

done it again, lost an hours work, can't even take a screenshot to save
my work. I can't type a rotate or rotate_extrude and hit preview. froze
up tight. mouse moves normal but no curser. What am I going to have to
do to not lose my work like this? Do a pulldown to "save-as" for every
line I type?

Cheers, Gene Heskett

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

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/gene

When I do a preview I see

Parsing design (AST generation)...
Saved backup file:
C:/Users/ChrisP/Documents/OpenSCAD/backups/potbox-backup-lzLPeAGl.scad

So if OpenSCAD were to crash I would open that file to get any unsaved
changes.

On Fri, 13 Aug 2021 at 13:28, Gene Heskett gheskett@shentel.net wrote:

Hey Gene!

General Linux stuff:  you can use this command to find any files in your
home driectory (~) or /tmp that were modified in the last 60 minutes:

find ~ /tmp -mmin -60

If you have a different mount point, you can add that between the "find"
and the "-mmin".  Modify Time is "less than", indicated by the "-", 60
minutes old.

Where are you saving your file?  Somewhere in your home directory?
Somewhere else on the local disk?  Somewhere on a remotely mounted volume:
NFS, SSHFS, etc?

Thus Gene Heskett hast written on Fri, Aug 13, 2021 at 08:28:45AM -0400, and, according to prophecy, it shall come to pass that:

On Friday 13 August 2021 10:58:05 Bryan Lee wrote:

Chuckle, but Bryan, I've been an mc user since my first red hat 5.0
install in 1998. I skipped the winders thing, learning just enough to
keep the first pc's we bought for the office help at WDTV where I was
the Chief for the last 19 years of my working life, coming up from
os9 on a color computer and amigados on a full blown amiga to red hat.

Old habits die hard. No winders allowed on the premises. But 6 linux
boxes run 24/7.

Today my back is demanding a decent chair, so I am logged into that
buster box from this stretch machine running the newest 07252021
openscad appimage on that machine and I'm wearing out the mouse doing
a save as every 2 or 3 characters. And I have arrived at something
that works as desired for a preview. Takes 2-3 seconds to show it
exactly like I want it to print on a Prusa mk3S.

But it doesn't render as anything but the first 2 of 4 items in a
difference scope. No advisories of any kind, my echo's look good,
but a F6 does not appear to get past the for loop that rotates the
last, effectively the third argument for the difference(). It says
the render done in just a fraction of a second. But it renders only
the first two lines of the differences scope.

Here is the .scad:
module armbearing()
{
// my feeble attempt at generating a ball bearing using bb's.
// NOW MODed for smaller drive, in ~/Dlds/3dp.stf/smaller-harmonic-drive
// armature is eccentric, 37.60mm min to 41.66mm max
// adjust india till it fits the armature when distorted by it.
// was 40.3 but in PETG fit is tighter
fg=.45; // some fudge for printeers over extrusion
vg=.20; //height of layer
india=38.385+fg; // in mm's of ID, inner should rotate with
// armature w/o walking
outdia=52.14-fg; // outer diameter, same as belt inner
height=6-vg; // in mm's, armature is 11.5 mm wide uses 2 of these
// everything below is generated center=true
// which means we're dealing in radius
ctr_ht=(height * .50000); // 1/2 height,mults are faster
inrad=india*.50000;  // rad of inner diameter
outrad=outdia*.5000;  // rad of outer diameter
bbtrkd=(outdia+india).5; // to center it
bbtrkr=(outrad+inrad).5; // ditto the radius
echo(198, inrad,outrad,bbtrkr,bbtrkd);
//adjust bearing size till it fits
bb=4.35+.2; // crosman bb so when inner is preloaded its about right. Was 4.45,
// slightly tight.
bbgap=.40; // wide enough to make inserting bb's easier
// how many bb's?
racircum=(inrad+outrad).50PI;//length of bb track
echo(205, PI, racircum);
bbs=racircum/bb;
echo(207,bb,bbs);
$fn=360;
// draw it
difference()
{
cylinder(h=height,d=outdia);// big outer for a base to sub everything else from
translate([0,0,-.05])cylinder(h=height+.1,d=india);// subtract inner
for(fn=[0:2:359]) // now cut ball track and insertion slot
{
rotate(fn)
{
union()
{
translate([bbtrkr,0,3])rotate([90,0,0])circle(d=bb,true);// subtract bb
// subtract bb insertion slot
translate([bbtrkr,0,3])rotate([90,0,0])scale([bbgap,1,1.1])square(height+1,true);
}
}
}
}
};

And two snapshots attached, one from the F5/preview, and one from the F6/render.
The first, preview is the desired result.
I see the last two translates and rotates could be moved to after the union()
on its line which might speed the render, but by how much?

Thanks for any clues as to why it doesn't render like it previews.

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

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/gene

You're using 2D shapes circle/square ? they fake render in F5, but without
a extrude they do not appear in 3D  / F6.

On Fri, 13 Aug 2021 at 20:08, Gene Heskett gheskett@shentel.net wrote:

On Friday 13 August 2021 14:26:20 Michael Möller wrote:

Replacements that work? rotate_extrude() can find more ways to refuse or
crash than websters has words. Most common is a seeming missmatch in
variable signs when their is not, even the error message showing the
allowable range has the same sign. But its bitching about that or any
one of the others. And doing it without identifying the the errant var
that caused the error.

wrote:

Cheers, Gene Heskett

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

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/gene

Rotate_extrude doesn't like singularities in the centre. You can't have
negative X coordinates but you can have a shape that hits the X axis but it
must be more than a single point, i.e an edge.

On Fri, 13 Aug 2021 at 19:46, Gene Heskett gheskett@shentel.net wrote:

On Friday 13 August 2021 15:28:35 nop head wrote:

With the exception of the originating cylinder with an ending true, there
are offsets in the 30mm range applied long before it gets anywhere near
the rotate_extrude(). Using rotate and a for loop I now have something
that may work, except the blue text has said 999/1000 now for a good 30
minutes. It finally got done but did not report the elapsed time. I made
a couple changes that should have sped it up, but didn't.

Is there a give up and show what it has timer?

wrote:

Cheers, Gene Heskett

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

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/gene

Gene: When you use a circle() as an object to do difference on a cylinder()
... you are subtracting a 2D object from a 3D object. "Does not Compute"
and crash bang. The preview does a nice thing "For Your Convinience" and
turns the circle into a 1mm cylinder. Thus the F5 looks good. The F6
however does things "By the book" and thus finds itself with conflicting
domain and hits a wall.

Do just a
circle(10);
and in F5 and F6 it shows the circle as a a cylinder or a"blue shadow",
respectively.

So, I've played with your code, and it will work (meaning F6 will render
the same as the F5 preview) if you change the circle to cylinder and square
to cube and give them - say - h=1 height. Rendering time isnt speedy,
admittedly.

On Fri, 13 Aug 2021 at 21:30, nop head nop.head@gmail.com wrote: