<othx>
InPhase linked to YouTube video "How to: Score and snap sheet plastic" => 1 IRC mentions
<InPhase>
ran_: It's just slightly stiffer with acrylic, but the same principle is used on the thickness you were looking at.
<ran_>
I get that but it doesn't help with drilling the mounting holes or the cutouts for the other hardware
ur5us_ has joined #openscad
<InPhase>
Well, sure, those would take some tools I guess.
<InPhase>
I guess if they're doing all the cutouts for that price, it's a fair deal.
<InPhase>
Skilled machine work labor costs a bit. You're paying for their experience.
ran_ has quit [Ping timeout: 252 seconds]
ran_ has joined #openscad
califax- has joined #openscad
califax has quit [Ping timeout: 276 seconds]
califax- is now known as califax
paddymahoney has quit [Ping timeout: 268 seconds]
paddymahoney has joined #openscad
<InPhase>
teepee: Got it narrowed down. I thought it might be this path, but I couldn't sort out quite how that did the recompute. But here's the trace for when it's happening: https://bpa.st/4C7A
<InPhase>
lookup_function to lookup_local_function to Context::create and then ScopeContext::init, over and over again.
<InPhase>
I was staring exactly at that last night when I got tired and gave up, so I must stare more at the same spot I guess until it clicks. :)
ran_ has quit [Quit: Leaving]
snakedLX has joined #openscad
snakedGT has quit [Ping timeout: 265 seconds]
<InPhase>
Okay. Half fixed! I have the recomputing disabled, but the top-level scope is not actually processed at all now, so next this must be inserted somewhere, presumably when the file is first inserted into the cache.
snakedGT has joined #openscad
snakedLX has quit [Ping timeout: 252 seconds]
<InPhase>
Yeah, this is the harder part. Nothing seems to be quite structured right to store that data at the right time. Tomorrow.
ferdna has quit [Quit: Leaving]
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
ur5us_ has quit [Ping timeout: 260 seconds]
califax has quit [Remote host closed the connection]
califax has joined #openscad
califax has quit [Client Quit]
califax has joined #openscad
califax has quit [Client Quit]
califax has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
Scopeuk has quit [Quit: Scopeuk]
Scopeuk has joined #openscad
lastrodamo has joined #openscad
dalias has quit [Ping timeout: 252 seconds]
dalias has joined #openscad
dalias has quit [Ping timeout: 260 seconds]
dalias has joined #openscad
mhroncok has joined #openscad
snakedLX has joined #openscad
snakedGT has quit [Ping timeout: 260 seconds]
Jack21 has joined #openscad
<Jack21>
Hi there - me again .. the "alt" (win) arrow or scroll function to increase or decrease is not working in addition or substractions like x=y-1
<Jack21>
x=y*1 is working
<Jack21>
probably something how it detects if there is something changable
arebil has joined #openscad
<InPhase>
Jack21: What if you put a space between - and 1?
<InPhase>
Jack21: That system is a little complicated and flaky in its implementation, relying on regular expressions instead of proper parsing. A bunch of the issues with it were patched up, but there are some that remain.
<InPhase>
y-1 might get parsed as -1 being an integer, but because y is adjacent it might get parsed as not an integer, thinking that y-1 is a variable.
<InPhase>
And I bet if you did "y -1" and scrolled up, it would turn to "y 0" and give you a syntax error...
arebil has quit [Ping timeout: 240 seconds]
<Jack21>
indeed a space works. with 2019.05 it was working without and y-1 became y+0
<Jack21>
also the cursor on the left ist not working to go 10 20 30 and if you have 10 and go down to 0 you can't go to -10 or back up to 10 as it is only 1 2 3 now
arebil has joined #openscad
<Jack21>
InPhase: so i can work with using spaces - but only advances by one or fractions is lame
<Jack21>
however now i end up with x=y+ +1; which is a bit awful , or when substracting it will then subtract a negative so adding it when using arrow down
Alexer has quit [Ping timeout: 252 seconds]
<InPhase>
Jack21: Yeah, between 2019.05 and 2021.01 one of our GSoC programmers was working on that while doing editor improvements, so there were changes. These were resolving some nasty malfunctions it used to do.
<InPhase>
Jack21: Overall I think it's not a great feature. I'm personally not even sure if we should keep it around.
Guest4356 has joined #openscad
<Guest4356>
hi
<Jack21>
changing a value and preview the result is a very useful feature
<InPhase>
Jack21: Alt and scroll has never worked on my system, because alt-scroll is used for window manager zoom, which is a hugely valuable feature that I'm not handing over to OpenSCAD. :)
Guest4356 has quit [Client Quit]
<InPhase>
Jack21: There's click and scroll, but this is a bad feature because it constantly changes which digit it's modifying, causing it to jump around inconsistently as you scroll.
Guest5 has joined #openscad
Guest5 has quit [Client Quit]
<InPhase>
Jack21: So I mostly find it a broken mess.
<Jack21>
the scroll function is useless but not the alt+arrow (not using the mouse when coding)
<Jack21>
i use this to find the correct value like you have a pow(i,y) in a for loop to distribute holes and you need the right value for y to reach the rim
<Jack21>
it is like a faster and more direct version of the customizer
<Jack21>
Also i use this in class for people (children) to understand better how the change of a variable changes the geometry. As you have fixed increments it is like a slow animation
Alexer has joined #openscad
lastrodamo has quit [Quit: Leaving]
raboof has quit [Ping timeout: 268 seconds]
raboof has joined #openscad
<InPhase>
Jack21: Well that's an excellent usage, and I can see the value there.
<InPhase>
And yes, the alt+arrow does tend to function better.
<InPhase>
Although you have to select carefully where you will use it.
<Jack21>
InPhase: or you eyeball a value and then tweak until the edge aligns - for most 3Dprints you only need 10μm precision
<Jack21>
when you need a tangential edge on a distorted sphere the math gets rather complicated
<teepee>
not really, we've disabled the test cases
<knielsen>
teepee: ok, then I'll do the same for Debian
<knielsen>
teepee: thanks
<teepee>
I don't think it's a good solution, but I don't see any way to fix that issue on our side
SamantazFox has quit [Ping timeout: 260 seconds]
<teepee>
I think it's a CGAL bug as the only way to prevent the crash would be to implement a different file reader validating the data
<knielsen>
teepee: so the problem is that CGAL does an assertion?
<knielsen>
at least that was mentioned in the cgal issue 5514
<teepee>
not sure if assert or just a crash, but it's in the stream operator, so there's no client side code
<knielsen>
right, a crash, that's not good, and agree it seems to be on CGAL to provide a way to catch invalid data without crashing
<teepee>
also std::cerr<<"wrong index in read_hedge"<<std::endl; in a library file reader is not something I would like to see
<teepee>
at least I think it crashes, but it's been a while since looking at that
<knielsen>
(fork() a process that does a trial read of the file, catch signal 6 or whatever from the child before doing real read... but not really, right?)
<teepee>
that would maybe help against crashing the application, but still keep the potential security issues, just in a different process
<dalias>
unshare and seccomp it :-p
<teepee>
not sure *if* it's actually security related still
<teepee>
yeah, we could send the file to some cloud service :)
<knielsen>
lol
<teepee>
well, I guess if someone is actually going to pressure this topic, it would probably just mean disable NEF for release versions
<teepee>
it's more a dev feature, I'm not sure any user is using this
<knielsen>
right
<teepee>
knielsen: did you see the FTBFS mail? I think the patch I've linked from the dev version should work fine for release too
<knielsen>
teepee: yes, that's what I wanted to fix now, and then I encountered the cgal test fails on the way
<knielsen>
hoping to get that FTBFS fixed quickly and look for a sponsor
<teepee>
ah, nice!
<knielsen>
and I should figure out how to get upload rights to Debian for openscad, there's a procedure to get a key signed by some other Debian devs which takes some work and I haven't really gotten round to yet
<InPhase>
Jack21: Most of my more designs have preview amounts that are too slow for incremental alt-arrowing to work out. But those sorts of adjustments can kick in as more valuable for demos, educational purposes, and simpler designs. The disadvantage to fine-tuning things in that manner though is that you lose parametric adjustment ability, which becomes more important at complexity scale. Hence it having a
<InPhase>
restricted valuable usage space. But definitely the one you're using it for is in that list.
<InPhase>
s/Most of my more designs/Most of my more serious designs/
<Jack21>
InPhase: I also have that problem (slowness) so my modules using a low fn when in preview and a higher for the render. Also i am trying to use polyhedrons as they are faster
<Jack21>
and some operations are only indicated but not executed - only in the render like minkowski stuff
<Jack21>
at last the alternative is to punch in the number press f5 and change the number again
<InPhase>
teepee: It's really a shame that the CGAL authors do not seem to value the "code should never crash" philosophy.
<InPhase>
Sitting here working on OpenSCAD with the "code should never crash" philosophy I end up pretty displeased with CGAL, when there aren't ways to make it avoid crashing.
<InPhase>
I guess the CGAL developers must be more math people than software engineer types?
<teepee>
yeah, in some cases we can solve it like peepsalot did with validating the stuff before, which is an acceptable way from library standpoint, specifically in regard to performance
<teepee>
but in this case there's no way to do input validation
<InPhase>
They have multiple 4 or 5 year old bugs for crashes, aborts, segfaults, and so on that are not getting developer attention.
<InPhase>
Okay... To be fair we have an 8 year old one, but it's a "cannot reproduce" hypothesized as related to an ancient OpenGL version, and should probably be closed. :)
<InPhase>
A 6 year old one which is marked as a crash, but is basically just that we don't have a way to exit processing while trying to close.
<InPhase>
Then our next oldest appears to be a crash in CGAL from 2015 that is still not fixed.
<InPhase>
It looks like a bunch of these older issues are dead issues though. Maybe someday we should clean up the ones that are probably not active anymore, like a #1623 with its random Windows-only crash that only happened on some systems after a driver update. Probably not our problem.
<InPhase>
The two reports in that issue are probably not even the same issue.
<Jack21>
i get oscad crashed regularly when opening 2 files where one is the lib of the other - normaly crashes when i start editing - doesn't matter if the change will or won't affect the other sheet
<InPhase>
Really?
<InPhase>
Like the program closes?
<InPhase>
That shouldn't happen, and I haven't seen it myself.
<Jack21>
yes it is so common that i am used to remember the line nr when starting editing so i can find the position faster when reopening it - Ü
<InPhase>
If you can do it reproducibly with the latest release or a nightly build, and can provide example code which does it, this is worth opening an issue.
<InPhase>
Provide steps as explicit as possible.
<InPhase>
That would probably be purely internal, so if you can get it to where we can reproduce it, it should be straightforward to fix.
<teepee>
if it's in openscad code :)
<Jack21>
i never could find why sometimes it is when changing file A sometimes when B .. but when i find something i let you know
<teepee>
not long ago I had crashes due to the editor component, but that does not seem to happen anymore
<InPhase>
Jack21: The tab additions were made by a young programmer who was still learning about RAII principles (if you know what those are), but I tried hard to review those and hammer out the instabilities in feedback. It's conceivable I missed one. So watch for a reproducible set of steps.
<Jack21>
last time it was with a warning - i clicked on it ↦ opened the lib and closed both when i started to fix the issue
<Jack21>
I also had this in 2019.05 without the tabs ..
<Jack21>
but i ll keep an eye on it - i had the feeling the 2021 version is more stable
<InPhase>
That's the goal. :)
<InPhase>
Robust never-crash code is a deeply held philosophy for me throughout my career, so these sorts of things will grab my attention when reproducible.
<InPhase>
And when it's properly fixable in our parts of the code...
<Jack21>
i could understand that a change after a save in the lib could cause something in the other file on execution - but it doesn't matter if you just enter a • within // comment text - i got the impression that it had something to do which file was opened first - but nothing that i could reproduce (but i didn't really try to - i tried to adapt my
<Jack21>
workflow so it doesn't happened)
<teepee>
hmm, it's not always possible to fix the bugs in a time frame that would be nice
<teepee>
but I can give a 100% guaranty we can't fix bugs we don't know about :)
<teepee>
really annoying are those bugs that can't be reproduced on the dev systems
<Jack21>
teepee: sometimes bugs get fixed when something else is updated or fixed without knowing about the bug
<Jack21>
so unintentional bug fixing exist
SebastianM has joined #openscad
<teepee>
yeah, true, it's not "fixing the bug" as specific action though ;-)
<Jack21>
unknown known is bad but unknown unknown is worse
<teepee>
indeed, that's where my coffee from amazon is right now
<Jack21>
are you referring to the coffee drone and the raven?
<teepee>
no, just a 1kg package of coffee beans: current status: "expected delivery 6. Sept."
<teepee>
it's in a place so unknown, even time stopped
<Jack21>
haha due to covid situations a lot of goods (70%) are not in a warehouse but stored on trucks on the road - leaving only 30% space for the actual goods to transport as the warehouses are full
<othx>
Jack21 linked to YouTube video "Australian Raven attacks drone delivering coffee in Canberra." => 1 IRC mentions
<teepee>
that raven has my full support :)
<Jack21>
if my neighbors start to order their coffee via drone delivery - ill start having/feeding ravens for sure.
<InPhase>
teepee: Amazon's chat tech support (works 24 hours a day) usually resolves those unknown past-due shipments quickly with a re-send attempt.
<InPhase>
teepee: They just hide the link for that well, and move it around often for fun, but it exists.
<teepee>
I know, I'm not too worried, they will likely just refund eventually
<InPhase>
I've never experienced them automatically refunding it unfortunately.
<teepee>
I've reordered meanwhile so I'm ok either way :)
<InPhase>
I think they assume by default that you got it and it didn't get scanned unless you complain.
<teepee>
oh? I had the same issue maybe 2 years ago and it just refunded maybe 3 or 4 weeks later
<Jack21>
last time i had an past-due - it turned out it was in the letterbox but without notification
<InPhase>
teepee: Maybe it differs where you are. But I had a lot of delivery chaos here that was not resolved, especially peaking in the last 2 years.
<teepee>
after starting with dhl packstation delivery it's pretty good, only gripe I have is that they split everything into maximum amount of tiny packages but it arrives without issues
<InPhase>
We had a President who thought he could become a supreme ruler by sabotaging our postal service, so amidst a pandemic making it already difficult, deliveries went pretty chaotic for a bit.
<InPhase>
The splitting is warehouse management. They don't keep all products in all warehouses.
<teepee>
ah, yes. the one you'll get back in 3 years
<InPhase>
I doubt it, but he seems to think so. Demographic shifts are moving strongly against him.
<Jack21>
i heard postal system in china is so good you can order your lunch in the morning by post
<InPhase>
Jack21: That probably only works in select cities. But we can also do that with uber eats. :)
<Jack21>
yes sure but i am not talking about delivery but the state postal service
<Jack21>
did you have seen these delivery robots already at your location (here they are probably not even allowed)
othx has quit [Ping timeout: 252 seconds]
<InPhase>
Never saw one.
<InPhase>
I can't imagine anything short of a flying drone surviving in Philadelphia though, which is right near me. Any land based bot would probably be crushed by car, attacked, or end up stuck in a meter deep construction pit.
<teepee>
worst case scenario... ran over by a tesla on "autopilot"
<InPhase>
I trust Teslas on autopilot about 50 times more than the average driver I deal with on my commute.
<InPhase>
I average witnessing more than 4 runnings of redlights and stoplights per hour. People also regularly stop in the middle of street resulting in wildly chaotic swerving into other lanes. I plan my commutes based on street width just so I can get around the crashes.
<teepee>
oh my, that sounds crazy
<teepee>
I guess I'm in a pretty quiet area and also don't drive much, so I'm maybe missing lots of what's going on in the middle of the city
SamantazFox has joined #openscad
<InPhase>
It's not the worst place I've ever commuted because there are a lot of alternate routes I can take, but it's probably the highest density of blatant disregarding of traffic rules that I've seen in my U.S. travels.
<InPhase>
Driving in Italy was a little worse. :)
<teepee>
ok, yes, very different style
snakedGT has joined #openscad
snakedLX has quit [Ping timeout: 265 seconds]
SebastianM has quit [Quit: Bye]
Polsaker has joined #openscad
<Jack21>
as one said - when we get fully autonomous transport/delivery we also will be in a fully surveillance area as only if everybody who kicks the robot can be sued and hold accountable
<Jack21>
else you will have people running on the street stopping cars just for fun - as there is no driver who kicks them for that
mhroncok has quit [Quit: Leaving.]
SamantazFox has quit [Remote host closed the connection]
SamantazFox has joined #openscad
snakedLX has joined #openscad
snakedGT has quit [Ping timeout: 265 seconds]
<teepee>
PSA: if you create a PR from your master branch, I might push updates into that :-)
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]