<linext_>
do you know whether it's possible to get stdout from a wasm as the prints occur?
<InPhase>
For computational benchmarking there is a recursive example024.scad (under Examples/Old) that is sometimes useful.
<InPhase>
But there are many computational things to test. There are also preview benchmarks that are different. It depends what you want to hit really.
<linext_>
i want to see how fast openscad wasm is likely to run. if the user has a slow computer, suggest offloading rendering to another user with a fast computer
<linext_>
so ex...024.scad takes 48 seconds running natively on my i9 9900
tachoknight has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
tachoknight has joined #openscad
chemhacker has quit [Quit: Leaving]
J1A8433 has quit [Quit: Client closed]
J1A8433 has joined #openscad
<InPhase>
Computational runtimes are best addressed with designs that don't render much geometry, if you want to do that part separately. Or designs that are done preview mode, but I think that is missing from the wasm approach if I recall correctly.
<InPhase>
But it depends what you want to pick out I suppose. Usually, for most designs, it's the CGAL runtimes that are most critical, so this is the key thing to test.
<InPhase>
(Unless you have reasons for otherwise, which could exist.)
vsellier has joined #openscad
JordanBrown has joined #openscad
tachoknight has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<JordanBrown>
Has anybody considered having animation cache the CSG trees for each frame, so that after one loop it could run faster? (But maybe they are too big for memory and reading them from disk would be too slow.)
<linext_>
animation could be multi-threaded
tachoknight has joined #openscad
<linext_>
are there examples of CLI outputting off, echo, ast, term ?
<JordanBrown>
I suppose that MT might help. I was just running an animation that (I believe) was execution-limited. It could calculate multiple frames in parallel.
<JordanBrown>
I don't think console logging was a relevant cost, because it's the same logging as for animations that run at full speed.
<JordanBrown>
Here I am referring to using animation in the GUI. If I was willing to drop to the CLI and could find my "make a movie" script, that would work, but isn't as good for iterative work.
<linext_>
so an "ast" file is flattened whitespace and some extra comments
<JordanBrown>
Sort of. It's actually decompiled.
<JordanBrown>
Note for instance that translate([10,10,10]) { cube(10); } displays in Design/Display AST without the braces.
tachoknight has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
la1yv_j has quit [Remote host closed the connection]
la1yv_j has joined #openscad
<teepee>
looks cool and also like a quite tough print
<J1A8433>
getting very thin at the top .. lets see what my printer does
<InPhase>
Tricky. Slowing printing will buy you cooling time, but might hurt bridging.
<InPhase>
Maybe print it upside down? :)
<InPhase>
Oh wait, that won't work at all.
<InPhase>
All those hexagonals...
<InPhase>
I guess that's a tricky one any way you slice it. Pun partially intended.
tachoknight has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
tachoknight has joined #openscad
<J1A8433>
hmm hole print is wobbling around .. probably need supports for the base center
<Friithian>
you're surprised that is wobbling?
linext__ has joined #openscad
ur5us has joined #openscad
linext_ has quit [Ping timeout: 268 seconds]
<J1A8433>
well that was a birdsnest after the 2 base rows of hexagons
qeed has quit [Read error: Connection reset by peer]
<J1A8433>
you really just cant't use the automatic supports
qeed has joined #openscad
J1A8433 has quit [Quit: Client closed]
J1A8433 has joined #openscad
teepee has quit [Ping timeout: 258 seconds]
teepee has joined #openscad
linext__ has quit [Quit: Leaving]
linext has joined #openscad
<peepsalot>
haven't been doing a lot of scad lately, but still getting somewhat frequent, yet unpredictable crashes here
<peepsalot>
for example just viewing some existing scripts and tweaking values
<peepsalot>
hmm, that's a new one, just got "ERROR: Parsing aborted by unknown exception" after enabling fast-csg, flushing caches, and previewing a script again. (script contains minkowski, so preview uses CGAL)
monkeybusiness has joined #openscad
<teepee>
I think it tries to be fully dynamic with toggling that, but I'm not sure it's 100% safe
qeed_ has joined #openscad
qeed has quit [Ping timeout: 250 seconds]
<InPhase>
peepsalot: Which platform?
<InPhase>
I don't recall seeing a crash in a long while.
<peepsalot>
linux mint 21
<InPhase>
Are these master builds?
<peepsalot>
nightly debian builds from obs, here is the exact script which seems to repeatably trigger the unknown exception message (not a crash) https://bpa.st/YCCQ
<peepsalot>
ah, yeah forgot about that. but also screenshot is convenient IMO to glance and copy settings
<teepee>
true, it's much easier to read
guerd87 has quit [Ping timeout: 244 seconds]
<InPhase>
peepsalot: "ERROR: Compilation aborted by exception: Unauthorized intersections of constraints"
<InPhase>
peepsalot: That's the .what() of the exception. I updated the handling of that to catch std::exception as a second-to-last resort.
<InPhase>
Also, one time I ran it with a boost format string error in my string format (oops), and this message popped up: https://bpa.st/JS6Q
<InPhase>
That sure looks a lot more like what I'd want to see. But I don't know why it only popped up that one time.
<InPhase>
Maybe it was pure coincidence that I had the boost format string error that run, because it didn't do it for the second boost format string error I made.
<InPhase>
But, I wisely saved that since it looked new and informative.