teepee has quit [Remote host closed the connection]
teepee has joined #openscad
<InPhase>
peepsalot: Bitten by incomplete standards declared as standard yet again.
<InPhase>
peepsalot: I was excited seeing the PR, then saw the result.
<InPhase>
peepsalot: Well, I suppose that's a fully integrated header-only library. You could always just add two conditions to the header to block filesystem support for those two platforms.
<InPhase>
By compiler version number perhaps.
<InPhase>
By the time anybody ever thinks to upgrade that header file, if ever, it probably won't be an issue anymore.
<peepsalot>
or i guess maybe its not intended as a macro expansion in boost, just the name of a template parameter
<JordanBrown[m]>
Is it a macro in the antique Bison being used?
<JordanBrown[m]>
However, the simplest thing might be to just move us away from using it. Replacing NE with, say. NEQ in parser.y and lexer.l should be safe.
fling has quit [Remote host closed the connection]
<JordanBrown[m]>
Three instances.
fling has joined #openscad
<peepsalot>
hmm, yeah, i guess that's easier than updating bison on the build server
fling has quit [Remote host closed the connection]
fling has joined #openscad
ur5us has joined #openscad
<peepsalot>
alright, mac is passing tests again!
epony has quit [Ping timeout: 272 seconds]
<JordanBrown[m]>
Sounds good!
<peepsalot>
now I just gotta figure out what happened to the windows build
<peepsalot>
aka issue #4408
fling has quit [Remote host closed the connection]
fling has joined #openscad
Guest25 has joined #openscad
Guest25 has quit [Client Quit]
fling has quit [Ping timeout: 255 seconds]
fling has joined #openscad
ur5us has quit [Ping timeout: 255 seconds]
epony has joined #openscad
epony has quit [Max SendQ exceeded]
epony has joined #openscad
epony has quit [Remote host closed the connection]
epony has joined #openscad
califax has quit [Ping timeout: 255 seconds]
califax_ has joined #openscad
califax_ is now known as califax
la1yv_a has joined #openscad
la1yv has quit [Ping timeout: 272 seconds]
la1yv_a is now known as la1yv
ur5us has joined #openscad
<Scopeuk>
it looks a little odd to me that we copy opengl32.dll from the mingq64 package if we are using mesa
paddymahoney has quit [Remote host closed the connection]
<Scopeuk>
from when I have played with mesa on windows over riding the system opengl with mesa opengl you (amongst other things) place the opengl dll in the same folder, I guess maybe our ming64/msys environment is build with a mesa opengl backend
<Scopeuk>
hmm ok that path is where the msys2 package puts the mesa dll
<Scopeuk>
having pulled the raw logs, the one that worked has an msys entry for installing mesa, the one that fell over does
<Scopeuk>
no ignore me I am blind
<Scopeuk>
why did that search not work
<Scopeuk>
git hub doing very strange things on those logs for the two link session when I ask for them
<Scopeuk>
thinking about it the recent versions of mesa I have used on windows did need mroe than 1 dll co locating with the executable
<Scopeuk>
and also appears to fail to correctly determine which backend to use and needed an environmental variable passing to make them go
<Scopeuk>
SET "GALLIUM_DRIVER=llvmpipe" was needed to force llvmpipe rather than having it "fall back" on its own which failed miserably
DenKn[m] has quit [Quit: You have been kicked for being idle]
aiyion has quit [Remote host closed the connection]
fling has quit [Ping timeout: 255 seconds]
aiyion has joined #openscad
fling has joined #openscad
snaked has quit [Ping timeout: 255 seconds]
fling has quit [Remote host closed the connection]
aiyion has quit [Remote host closed the connection]
paddymahoney has joined #openscad
fling has joined #openscad
aiyion has joined #openscad
<Scopeuk>
its a bit of a fudge but I pulled down the various msys packages that supply mesa and it's dependencies, dumped all the dll's in with my snapshot exe and it does crash on launch
fling has quit [Remote host closed the connection]
fling has joined #openscad
<Scopeuk>
hmm this appears over dead
Guest69 has joined #openscad
Guest69 has quit [Client Quit]
Sauvin has quit [Quit: Leaving]
Sauvin has joined #openscad
paddymahoney has quit [Remote host closed the connection]
teepee has quit [Ping timeout: 255 seconds]
teepee has joined #openscad
aiyion has quit [Remote host closed the connection]
<Scopeuk>
peepsalot I started and gave up a little when the massive stack of dependencies went in a different direction. it worked fine using the direct release from https://github.com/pal1000/mesa-dist-win
<Scopeuk>
not sure if that is the libraries or how I assembled the runtime enviroment
<Scopeuk>
I guess we could try using pinned version of mesa and it's support components
<Scopeuk>
ahh looks like you are ahead of me on that
<peepsalot>
yeah, no worries. thanks for looking into it anyways.
<peepsalot>
should find out in a few minutes if that was the issue.
fling_ has joined #openscad
fling has quit [Ping timeout: 255 seconds]
<Scopeuk>
I do wonder why msys appear to use a random compression system invented by and seemingly only otherwise used by facebook
<peepsalot>
i think the underlying package manager, pacman, is using zstd
fling_ is now known as fling
<peepsalot>
pacman is what Arch uses
<Scopeuk>
ahh, I'm not familiar with it, some of their packages are still .tar.gz
<Scopeuk>
yeh my experience is mostly with debian derivitives
<Scopeuk>
hmm windows run still doesn't look happy
fling has quit [Remote host closed the connection]
<peepsalot>
hmm, failed only 1103 of 1520 tests, rather than 1175 of 1520
fling has joined #openscad
<peepsalot>
so at least a small number of the graphics tests passed somehow
<Scopeuk>
looks like the same illegal access return code as before
<Scopeuk>
when I was playing with the local libraries their ver difference in the dependencies for mesa between 22.1.7 and 22.2.1
<teepee>
Friithian: you said 22.04 which implies ubuntu and OBS has ubuntu packages ready to just apt update && apt upgrade, or even unattended-upgrades if you do that
<peepsalot>
so i'm thinking of seeing if those two changes would help
<teepee>
I think the msys action caches stuff somewhere, but I'm not exactly sure about the behavior of that update flag
<Friithian>
teepee: my workflow already have building from source in it! Even though something's going wrong
<teepee>
that's fine of course :)
<teepee>
oh, actually I need to check if they already provide the latest ubuntu release
<Friithian>
and what's breaking is kinda confusing me but I dont know the codebase at all so *shrug*
<Friithian>
oh it isn't just my system being fuckey?
<teepee>
nope
<Friithian>
wow
<Lagopus>
Is it possible in openscad to perform nonlinear transformations/skews/extrusions? (I have assumed the answer is no.)
<Friithian>
is there an issue already open on this or do I need to open one?
<teepee>
Lagopus: not with multmatrix
<teepee>
Friithian: I don't think we need a separate issue, it's related to the currently ongoing changes, so it's pretty much work-in-progress and I'm certainly aware of all the red marks at https://build.opensuse.org/package/show/home:t-paul/OpenSCAD ;-)
<Friithian>
ok cool cool, time to git switch :D
<teepee>
I just need to make a bit progress on the christmas gift prints, then I want to have a look at that
<Lagopus>
teepee: I imagine I will need to do point-per-point calculations of the existing surface to create the next surface to 'extrude' to
ur5us has quit [Ping timeout: 246 seconds]
<Lagopus>
I appreciate the quick answer, thanks.
<teepee>
Lagopus: yes, right now there's pretty much the only option to keep the points in user space
<teepee>
BOSL2 is doing that, it might also have some of the stuff you are looking for
<Lagopus>
I will check BOSL2 out. Thank you again.
To_Aru_Shiroi_Ne has joined #openscad
ToAruShiroiNeko has quit [Ping timeout: 264 seconds]
To_Aru_Shiroi_Ne is now known as ToAruShiroiNeko
J1A84 has quit [Quit: Client closed]
J1A84 has joined #openscad
<peepsalot>
Friithian: I didn't realize the OBS stuff was still broken, but I think SpaceNavInputDriver.cc is just missing: #include "printutils.h" at the top, if you want to try it real quick
<Friithian>
ill try that
<teepee>
peepsalot: the printdb has %, are those comma now?
<teepee>
I seem to remember some discussion regarding that boost format stuff
<Friithian>
come on clangd prase include faster u fuk
<Friithian>
oh that seems to have fixed it
<peepsalot>
teepee: i don't think so, I only changed the STR macro, PRINTDB is a completely independent weird macro
<Friithian>
peepsalot: yup that works, thanks
<teepee>
ah, ok, also confirmed by the include fixing the issue
<teepee>
Friithian: chance for a nice little PR :)
<Friithian>
…
<Friithian>
well I guess I'm pausing whatever I was doing (which I've already forgot lma) to make a PR
<teepee>
Friithian: just edit directly on the github website
<Friithian>
oh yeah I keep forgetting github has that
<teepee>
the edit button will automatically generate a PR for the change
<teepee>
"Build and test" section: 100% tests passed, 0 tests failed out of 1545 :)
<peepsalot>
cool, i got UCRT64 version to build, but it does nothing to help the GL issue
la1yv has quit [Read error: Connection reset by peer]
la1yv has joined #openscad
<peepsalot>
the ucrt is something that is part of Win 10 and up. but windows 7 etc can possibly download and install it on their own. I kinda wonder if this would actually help with the compatibility issues we were seeing
<teepee>
unfortunately not, that's all msys2 right? the release build causing the issue with win7 is the MXE build
<teepee>
not sure if there would be a way to package the msys2 binaries
<peepsalot>
ah, right, I forgot the releases used MXE
<Friithian>
hnnng with the 5900x openscad built in 2 minutes 2 seconds
<peepsalot>
sounds slow :P
<peepsalot>
[100%] Built target OpenSCAD real 1m42.335s
<Friithian>
this was a build from nothing, no cache or anything (and I was updating system packages at the same time)
<peepsalot>
on my 3900X
<peepsalot>
i wiped build dir and started from scratch also
<teepee>
nice, first ubuntu builds are green on OBS \o/
<Friithian>
does that include wiping and ccache?
<Friithian>
ssdsys
<Friithian>
because I just did `make clean && time make -j24` and it ``built'' in 0m11.308s
<Friithian>
(but hey that means it's caching really well)
ur5us has joined #openscad
<Friithian>
teepee: macos is good too, right?
<teepee>
in general, no ;-). what do you mean specifically?
<teepee>
the build fix? that code is linux only
<Friithian>
oh just wondering if it was just ubuntu green or only windows red
<teepee>
windows is the issue peepsalot is fighting with
<teepee>
the spnav driver only runs on linux or freebsd as far as I know
<Friithian>
``Helplessly flailing in the general direction of Windows CI'' lol yeah
<teepee>
yeah, looks like something broke the opengl driver stuff :(
<Friithian>
bad solution: vulkan :D
<teepee>
vulkan would be awesome, but would that work on plain windows?
<Friithian>
I think so
<teepee>
but of course that will make it more complicated on macos
<InPhase>
Demangled but not comprehensible. Thanks C++
<InPhase>
But clearly a template mess. :)
<peepsalot>
"string table overflow at offset 10000327" coming from assembler, sounds like a limitation of the [whatever executable/debug format], but i have no idea.
snaked has quit [Ping timeout: 255 seconds]
<peepsalot>
if that's hex, then that would be just over 256MB
<peepsalot>
also i'm not sure if there's any point to making a debug build without using interactive debugger...
<peepsalot>
but I think I remember debug builds sometimes spitting out a little extra info about the code unit that crashed (but not a full stack trace)?
<peepsalot>
i could just be making things up in my head, and it all may be moot anyways if it won't build in the first place.