<teepee>
a bit more setup and this really starts to be usable :)
<OlivierChafik[m]>
I didn't do anything!!
<OlivierChafik[m]>
try and refresh
<OlivierChafik[m]>
I think the service worker i use to trick the browser into allowing SharedArrayBuffer only works after a reload (on Chrome it works without it because I asked for an exemption). Maybe your SW was down or something. You can also turn it down in the Application / Service Workers and reload twice to see
<teepee>
getting now EXCEPTION: ReferenceError: SharedArrayBuffer is not defined
<OlivierChafik[m]>
teepee: weird. If i run it from a private browsing window it fails all the time. If i load it from my main profile it fails the first time after I killed the service worker, then works after a refresh
<OlivierChafik[m]>
something about private window + service worker behaviour. Now I hate firefox thanks.
<OlivierChafik[m]>
(ugh and it seems you can only debug service workers with the nightly builds of firefox, right)
redlizard has joined #openscad
<myosotis>
service workers made my head hurt. so buggy and different between browsers. I gave up on every project I tried it on maybe 2ish years ago
LordOfBikes has quit [Ping timeout: 256 seconds]
LordOfBikes has joined #openscad
myosotis has quit [Quit: myosotis]
drkow has joined #openscad
snaked has quit [Read error: Connection reset by peer]
snakedGT has joined #openscad
drfff has quit [Ping timeout: 256 seconds]
ali1234 has quit [Remote host closed the connection]
ali12341 has joined #openscad
J2269 has joined #openscad
ur5us has quit [Ping timeout: 240 seconds]
J22 has quit [Ping timeout: 256 seconds]
luissen has quit [Ping timeout: 250 seconds]
ferdna has quit [Quit: Leaving]
teepee_ has joined #openscad
GNUmoon has quit [Ping timeout: 240 seconds]
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
GNUmoon has joined #openscad
king has joined #openscad
GNUmoon has quit [Ping timeout: 240 seconds]
king has quit [Quit: Client closed]
GNUmoon has joined #openscad
ur5us has joined #openscad
ur5us has quit [Ping timeout: 240 seconds]
gunnbr_ has joined #openscad
othx has quit [Read error: Connection reset by peer]
othx has joined #openscad
gunnbr has quit [Ping timeout: 256 seconds]
arebil has joined #openscad
KimK has joined #openscad
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
arebil has joined #openscad
foul_owl has quit [Ping timeout: 256 seconds]
foul_owl has joined #openscad
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
<teepee>
that was needed for lib3mf enabling the MINGW stuff
<peepsalot>
the way that line is defined it treats it like a single definition in a weird way info.cmake shows DEFINES = ... ENABLE_CAIRO;__GCC -DENABLE_LIB3MF;USE_QOPENGLWIDGET; ...
<teepee>
hmm, yeah, looks like it's 1 string and probably gets split later, that's a bit messy
<teepee>
it's a normal define IIRC
<teepee>
so those 2 should be separate -D
<peepsalot>
and then there ends up being a line in moc_predefs.h: #define __GCC -DENABLE_LIB3MF 1
<peepsalot>
whatever that does
<teepee>
ouchies
<dalias>
that sounds very broken
<teepee>
well, it defines __GCC with that string
<teepee>
which still counts as #isdef
<dalias>
well yes, but ... defining macros in private __GCC namespace ??
<teepee>
but it's certainly not that it should be :)
<peepsalot>
i mean, i get what the define does, i don't know what that generated header does though
<teepee>
probably just some crazy Qt stuff, it's not needed there
<teepee>
it's really just the nonstandard use of defines in lib3mf that needed it
<teepee>
otherwise it would break building the stream stuff with the VS code path
<OlivierChafik[m]>
We'll most need something similar to the change mentioned at the end. Although I'd rather have wider intervals + still filtering than just no filtering, for speed reasons.
<OlivierChafik[m]>
I used to think (before reading that bug) we could switch all of corefinement to Cartesian<Gmpq>, which I tried (to be consistent w/ the legacy nef) and is slower. But that wouldn't help with wasm if it still uses epeck inside.
<OlivierChafik[m]>
teepee We'll most likely need something similar to the change mentioned at the end. Although I'd rather have wider intervals + still filtering than just no filtering, for speed reasons.
<OlivierChafik[m]>
I used to think (before reading that bug) we could switch all of corefinement to Cartesian<Gmpq>, which I tried (to be consistent w/ the legacy nef) and is slower. But that wouldn't help with wasm if it still uses epeck inside.
<teepee>
it's a bit confusing as the PR mentions "cool, now we can support platforms without rounding-math"
<OlivierChafik[m]>
(I hope to have more time later this week to test modified kernels + patch things up)
<teepee>
but it seems like 96,29% broken
<OlivierChafik[m]>
> the PR mentions "cool, now we can support platforms without rounding-math"
<teepee>
and it's referencing valgrind instead of wasm, but the issue with valgrind is the same that it does not support rounding-math
<OlivierChafik[m]>
ah yes
<teepee>
and the test they show (in the PR I think) with a big shere and lots of small spheres cut out does not work with any random resolution I tried