RichardPotthoff has quit [Ping timeout: 245 seconds]
mmu_man has joined #openscad
sculptor has joined #openscad
howiemnt1 has quit [Ping timeout: 276 seconds]
howiemnt has joined #openscad
church_ has quit [Remote host closed the connection]
church_ has joined #openscad
j45_ has joined #openscad
j45 has quit [Ping timeout: 276 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #openscad
mmu_man has quit [Ping timeout: 244 seconds]
L29Ah has joined #openscad
mmu_man has joined #openscad
sculptor has quit [Ping timeout: 244 seconds]
califax has quit [Remote host closed the connection]
califax has joined #openscad
pah is now known as pa
mmu_man has quit [Ping timeout: 244 seconds]
jbd has joined #openscad
feuerrot has quit [Server closed connection]
feuerrot has joined #openscad
sculptor has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
mmu_man has joined #openscad
LordOfBikes has joined #openscad
NoGodDamnIdea has joined #openscad
shwouchk has quit [Quit: WeeChat 4.5.2]
mmu_man has quit [Ping timeout: 255 seconds]
mmu_man has joined #openscad
JordanBrown has joined #openscad
<JordanBrown>
I have a dim memory that there has been a recent Github discussion that touched on dynamically downloading components, either as a part of the cmake process or as a part of the make process itself. Was that real or a product of my fevered imagination?
<JordanBrown>
That discussion about closing stale bugs has prompted me to look into #331. I think the fix is pretty easy.
<teepee>
providing all the native linux distro builds, currently for Raspbian, Debian and Ubuntu
<JordanBrown>
Cool.
<JordanBrown>
Right now, if you have a string with an unknown backslash escape (including a malformed \x or \u), the result is that the backslash is ignored and emitted to stdout(!).
<teepee>
just looking at the sphere comment
<JordanBrown>
Not trying to do work has left me with a lot of time :-)
<teepee>
I think it might be better to have a dedicated primitive, like blender has icosphere in addition
<JordanBrown>
How many ways are there to make spheres? Do we really want new words for each of them?
<teepee>
ah, putting it on stdout seems wrong :)
<JordanBrown>
The only reason that I made it an option was that *somebody* probably depends on the current mesh.
<JordanBrown>
(And, come to think of it, I should check how it matches up with a cylinder.)
<teepee>
yes, so the current sphere should probably stick around
<JordanBrown>
The octahedral subdivision might well match up, for multiples of 4 sides.
mmu_man has joined #openscad
<JordanBrown>
But *somebody* will depend on a sphere having flat poles, or depend on matching up with an odd number of sides.
<JordanBrown>
One other thought would be to have $spherestyle, which would be used sphere to decide which style to use. You could set it on individual calls, or you could set it globally.
<JordanBrown>
And that would then provide a way to transition to a better scheme: it could first default to "orig", and adventurous people could set it to "octa". It could later default to "octa", and people with obscure legacy could set it to "orig".
<JordanBrown>
Has anybody else seen a behavior on recent builds where there is a couple-of-second delay on startup?
<JordanBrown>
I'm seeing either a white window for a couple of seconds, or a normal window that is then locked up for a few seconds apparently running the empty program.
<JordanBrown>
I'm on a fairly low-powered laptop, so others on higher-end platforms may not have it as visible.
<JordanBrown>
I first noticed it on my own build, but it's there on the snapshot too.
<JordanBrown>
It's not there in 2021.01, and it is there in 2025.08.14.
<JordanBrown>
2021.01 is maybe 2s from clicking [New] until it's ready.
<JordanBrown>
In my most recent attempt, 2025.08.14 was around 17s.
<JordanBrown>
This may be related to the "pull down the print services" feature and my weak network connection, but I think it was there when I had *no* network connection.
<JordanBrown>
And I thought that I had that feature disabled.
<JordanBrown>
But maybe I misremember how to disable it.
<JordanBrown>
Back on #331... the question then arises: what *should* happen when you have an undefined or malformed escape sequence?
<JordanBrown>
a) drop the \, retain the rest
<JordanBrown>
b) retain all of it as normal text
<JordanBrown>
c) error
<JordanBrown>
d) something else
<JordanBrown>
My inclination is (c) error.
<JordanBrown>
That way you can't depend on undefined behavior that might change in the future.
<JordanBrown>
I also notice some ... surprising ... behavior in the immediate vicinity:
<JordanBrown>
* a LF (newline) in the middle of the string is ignored, so the string continues on the next line.
<JordanBrown>
* a CR in the middle of the string is copied intact.
<teepee>
the delay is probably the font update, I'll try to find a different way, but Qt is ignoring some of the initial setup unfortunately
<JordanBrown>
Doesn't pop the fontconfig dialog.
<JordanBrown>
There's also the weird fact that *sometimes* it seems to run the empty program, and sometimes not.
<teepee>
nothing to do with that, it's the "set default GUI font"
<JordanBrown>
OK. I don't know anything about that.
<teepee>
that it's executing the script is a separate issue I think
<JordanBrown>
At least on this low-powered laptop, it's a bit of a nuisance.
<JordanBrown>
I just reread what you wrote. I agree that on its face it seems like the white-screen delay and the running-the-empty-program delay seem separate, except...
<teepee>
I'm seeing that too as I usually run my notebook in power saver mode which is great, no fan noise ever :)
<JordanBrown>
that they seem to be exclusive. White screen means no running-empty-program, and vice versa.
<teepee>
well, Qt setStyle() and running the program having a direct connection does not make sense
<teepee>
it could be that the run-issue has some time/race-condition attached so the delay triggers it more often
<JordanBrown>
No it doesn't. But the pattern remains :-)
<JordanBrown>
doesn't make sense, that is
<JordanBrown>
I'll spend an hour or two trying to nail down where the delay is.
<JordanBrown>
I can live with a style checker, but I'd really prefer one that said what it was unhappy about, not just "code should be clang-formatted".
<teepee>
ideally it should create a patch and post that on the PR
<JordanBrown>
are you saying that it *does*, or that you think that would be best?
<teepee>
I'm not sure what it does exactly right now
<JordanBrown>
In working with a style checker at work for an appalling number of years, I find that often my idea of good style and its idea are different, and so I try to come up with the variation that I consider to be the least-bad variation that makes it happy.
<JordanBrown>
That's not always possible, of course, so sometimes I have to hold my nose.
<JordanBrown>
It thinks, for instance, that {1+2} should be formatted as {1 + 2}. That is, it wants spaces around the plus, but not inside the braces.
<JordanBrown>
I think that that visually means that the open brace and 1 bind together tighter than the 1 and the 2, and of course they don't.
howiemnt1 has joined #openscad
marcus has quit [Server closed connection]
marcus has joined #openscad
howiemnt has quit [Read error: Connection reset by peer]
<JordanBrown>
Right now, I consider this checker unusable because in some cases I can't figure out what it's complaining about, after trying several variations.
<teepee>
it does upload the patch as artifact
<teepee>
using that is not working?
<JordanBrown>
Ah, I have not found that yet.
<teepee>
and, argh! it seems to have exactly that issue with line length :(
<teepee>
which is taking about 50% of the screen on my 13" notebook
<teepee>
increasing the unused space even further
howiemnt1 has quit [Remote host closed the connection]
<JordanBrown>
I actually prefer that one as five lines.
howiemnt1 has joined #openscad
<JordanBrown>
But thanks for pointing me at the patch; I'll just blindly do what it says.
<JordanBrown>
I prefer that the three expressions each be on their own lines, so that the fact that there are three similar expressions is very visible.
<teepee>
I guess for such cases the only option is some sort of annotation to keep it as it is
<teepee>
no auto format I have seen gets >90% of formatting that makes sense
<JordanBrown>
The one at work has such an annotation, including (at my request many years ago) a block form of the annotation.
<JordanBrown>
But in general I find that adding a per-line annotation introduces more ugliness than just doing what it wants does.
<JordanBrown>
What is the line length set to? For that spherical_to_xyz case it seems to be formatting for well more than 80 columns.
<teepee>
it seems to break at about 100 or so
<teepee>
ColumnLimit: 105
<teepee>
meh
<teepee>
:(
<JordanBrown>
I have a mild preference for 80 just for legacy reasons and to maintain consistency across a broad range of work, but I do think that *some* limit is appropriate.
<teepee>
*some* limit is fine if it's >= 200
<teepee>
even on my small 13" screen I would like to use at some of the screen space
<JordanBrown>
That would be a lot for me even at home where I have ~80 horizontal inches of screen.
<teepee>
and 80 column expression breaks are in 98.35% of all cases just horrible
<JordanBrown>
On a 15" screen, it's too much. I like to be able to see something *else* on the screen.
<teepee>
I'm not saying use it to set the code in block formatting
<teepee>
and if need be use variables to break things apart
<JordanBrown>
In primitives.h, the fact that it wants to collapse all of those parallel << operations is just awful.
<teepee>
there might be a setting for that
<JordanBrown>
I'm not understanding what you mean by "not saying to use it to set the code in block formatting".
<JordanBrown>
It sure seems to want to wrap there, perhaps tempered a bit by semantic units.
<teepee>
like a book, have all lines that long except the last one
<teepee>
I don't know what that's called in english
pca006132 has quit [Server closed connection]
pca006132 has joined #openscad
<JordanBrown>
Assuming that we aren't talking about wrapping multiple statements together, right, that seems to be what it wants.