teepee changed the topic of #openscad to: OpenSCAD - The Programmers Solid 3D CAD Modeller | This channel is logged! | Website: http://www.openscad.org/ | FAQ: https://goo.gl/pcT7y3 | Request features / report bugs: https://goo.gl/lj0JRI | Tutorial: https://bit.ly/37P6z0B | Books: https://bit.ly/3xlLcQq | FOSDEM 2020: https://bit.ly/35xZGy6 | Logs: https://bit.ly/32MfbH5
linext has quit [Quit: Leaving]
dalias has joined #openscad
dalias has quit [Ping timeout: 276 seconds]
<peeps> pca006132: RelWithDebInfo i think does higher optimization than -Og by default, not sure which one wins out when added to CMAKE_CXX_FLAGS (whichever shows up last on the compiler command line typically)
<peeps> for a pure CMAKE_BUILD_TYPE=Debug build however, it seems to default to no optimizations at all, so the -Og helps there
<peeps> -Og and all other optimizations are described here: https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html
<peeps> in the past when I was doing a lot more stack tracing and profiling, i also found -fno-omit-frame-pointer seemed to help with more complete traces in some cases
peeps is now known as peeps[zen]
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
dalias has joined #openscad
fling has quit [Ping timeout: 240 seconds]
dalias has quit [Ping timeout: 255 seconds]
LordOfBikes has quit [Ping timeout: 255 seconds]
fling has joined #openscad
misterfish has joined #openscad
fling has quit [Remote host closed the connection]
fling has joined #openscad
LordOfBikes has joined #openscad
qeed has quit [Quit: qeed]
qeed has joined #openscad
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
kintel has joined #openscad
misterfish has quit [Ping timeout: 264 seconds]
<gbruno> [github] kintel pushed 1 modifications (Bug candidate 2a) https://github.com/openscad/openscad/commit/981ca840c11f8e33a4f1b917ab6e66bb7cdfea58
<gbruno> [github] kintel synchronize pull request #4840 (Allow building without CGAL) https://github.com/openscad/openscad/pull/4840
dalias has joined #openscad
dalias has quit [Ping timeout: 264 seconds]
J23k87 has joined #openscad
J23k15 has quit [Ping timeout: 250 seconds]
<pca006132> I think CMKAE_CXX_FLAGS wins, at least it linked properly
<pca006132> yes, -fno-omit-frame-pointer can sometimes help
<pca006132> the smart pointer thing is more about maintainability
pca006132 has quit [Remote host closed the connection]
pca006132 has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mmu_man has quit [Ping timeout: 256 seconds]
pca006132 has quit [Remote host closed the connection]
Guest50 has joined #openscad
Guest50 has quit [Quit: Client closed]
pca006132 has joined #openscad
dalias has joined #openscad
<gbruno> [github] pca006132 opened issue #4858 ( VBO renderer uses excessive RAM on Windows) https://github.com/openscad/openscad/issues/4858
<pca006132> it turns out the OOM issue I got yesterday is not related to the indexed polyset patch, but something I can reproduce on current master
<pca006132> it is doing something like 1kb per vertex, which quickly explodes the memory
pca006132 has quit [Remote host closed the connection]
TheCoffeMaker has quit [Ping timeout: 256 seconds]
TheCoffeMaker has joined #openscad
dalias has quit [Ping timeout: 255 seconds]
misterfish has joined #openscad
misterfish has quit [Ping timeout: 255 seconds]
pca006132 has joined #openscad
lastrodamo has joined #openscad
dalias has joined #openscad
dalias has quit [Ping timeout: 255 seconds]
dalias has joined #openscad
dalias has quit [Ping timeout: 260 seconds]
arebil has joined #openscad
misterfish has joined #openscad
lastrodamo has quit [Quit: Leaving]
kintel has joined #openscad
<gbruno> [github] gsohler synchronize pull request #4842 (Indexed2 squash) https://github.com/openscad/openscad/pull/4842
<kintel> pca006132 Ouch, worth opening a ticket perhaps? I assume the 1kb comes from CGAL/Gmp
<kintel> ..or perhaps mimalloc is misbehaving somehow
<pca006132> it seems that it is from some OpenGL stuff
<pca006132> tried disabling mimalloc but it does not make a difference
<pca006132> I would be great if someone with a physical machine can test this, to rule out potential issue from mesa3d on windows
<kintel> Thx!
<kintel> Well, perhaps time has come for me to try successively reinstalling Windows components until I can build again : (
<pca006132> I feel like this might be due to mesa because otherwise someone should have already complained...
<pca006132> btw, you should run it under gdb and prepare to interrupt it when you see the RAM spike...
<pca006132> or you will get OOM in 1s
dalias has joined #openscad
<kintel> pca006132 it just sits expanding memory use after F6, without any user interaction?
<kintel> ..yet it seems to keep calling rendering-related functions when it should not be redrawing?
mmu_man has joined #openscad
dalias has quit [Ping timeout: 256 seconds]
<kintel> ..and this is both with/without manifold enabled as well?
<pca006132> no, it immediately consume all the memory after you hit render
<pca006132> it doesn't even finish the F6 render
<kintel> Does it fail in cmd-line mode too, or just GUI?
<pca006132> just GUI
<pca006132> I think it is with manifold enabled, but I don't think it is related
<pca006132> for memory usage increase by about 1kb per vertex, I added a breakpoint to the VertexArray::createVertex function and was stepping through it
<pca006132> when I reach the end of the function gdb said the heap was corrupted so it cannot be stepped. When I do a continue the program is trapped in another call to createVertex and the memory usage increased by about 1kb
<kintel> pca006132 Did you try the latest binary snapshot too? ..just to focus to problem to self-built binaries
<kintel> I'll try fixing up my Windows box, but it may take a day or two to get anywhere
<pca006132> no, forgot to try that
<pca006132> will try
<pca006132> let me finish up with the manifold fuzzer thing first
<kintel> ..but your diagnosis should allow me to look for problems by reading code too. I just want to have a good repro case :)
<kintel> At least I know how the vertex array code works now. For some definition of "works"
<kintel> Also, trying this CL would be cool: https://github.com/openscad/openscad/pull/4789
<kintel> I haven't rebased it to master in a while, but it should kill any non-VBO code paths and clean up some of the VBO stuff
guso78k has quit [Quit: Client closed]
<pca006132> sure, will try
guso78k has joined #openscad
<guso78k> pca006132 the PR is green now!
<guso78k> thank you for spotting the bug. the bad alloc was there for good. there was too less memory available the the facenum coded as Ascii text :)
<guso78k> i am just wondering which tools did you use to catch the bug.
<pca006132> address sanitizer
<pca006132> it will dump a trace when this kind of things happen
<guso78k> did the same, but did not flag any error. strange
<guso78k> again, thanks!!!!!
<guso78k> i think i can still implement some of your comments. maybe i have time this evening
<pca006132> kintel: yes it also OOM with the prebuilt binary OpenSCAD-2023.11.24 x86 (64-bit) - zip package
<pca006132> it works fine with #4789
<kintel> pca006132 Cool perhaps I inadvertently fixed a bug :)
<pca006132> yeah this is nice
<pca006132> perhaps some legacy code path thing
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
misterfish has quit [Ping timeout: 276 seconds]
L29Ah has quit [Read error: Connection reset by peer]
L29Ah has joined #openscad
mmu_man has quit [Ping timeout: 256 seconds]
misterfish has joined #openscad
aiyion1 has quit [Remote host closed the connection]
aiyion1 has joined #openscad
mmu_man has joined #openscad
aiyion1 has quit [Remote host closed the connection]
aiyion1 has joined #openscad
mmu_man has quit [Ping timeout: 276 seconds]
mmu_man has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
TheAssassin has quit [Remote host closed the connection]
TheAssassin has joined #openscad
TheAssassin has quit [Remote host closed the connection]
TheAssassin has joined #openscad
RHino_1 has joined #openscad
RHino_1 has quit [Client Quit]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
kintel has joined #openscad
<kintel> pca006132 FYI: It doesn't OOM on nvidia drivers
<kintel> Managed to build on Windows again, but I have to build as Release, otherwise the compiler barfs
pca006132_ has joined #openscad
<pca006132_> so it is probably something with mesa
<pca006132_> if you try -CMAKE_CXX_FLAGS="-Og" you should be able to get a RelWithDebInfo build
<pca006132_> but as this is fixed by the VBO refactoring and only affects software renderer, I guess it is not that important...
pca006132_ has quit [Quit: Client closed]
pca98 has joined #openscad
pca98 has quit [Client Quit]
zauberfisch has quit []
<gbruno> [github] JCU0 opened issue #4859 (STL import. Easy (I hope) fix for degenerated triangles.) https://github.com/openscad/openscad/issues/4859
LordOfBikes has quit [Remote host closed the connection]
dalias has joined #openscad
dalias has quit [Ping timeout: 268 seconds]
pca006132_ has joined #openscad
pca006132_ has quit [Client Quit]
<teepee> J23k87: starting with a little printable deko - https://imgur.com/a/1MxKz7q ?
<guso78k> use the lighter slightly to remove the stringing and your deco will look perfect
<teepee> it's not so much about the model itself, but if we can get an advent calender 2023 :)
<teepee> it's like waaaaaay too late again
<teepee> as every year
<J23k87> teepee  i see your nice new printer
<teepee> yes, that thing is pretty amazing
<teepee> it looks quite similar to the ancient i3 MK1 but the details are miles ahead
<J23k87> afaik the upgrade kit from 3S to MKⅣ is quite expensive
<teepee> yeah, full kit would have new 32bit electronics + all new motors (0.9° steppers)
<J23k87> oh really that sound nice
<teepee> also the new extruder I believe which is quite a cool idea, using the nozzle directly for bed leveling
<teepee> no extra sensor that needs to be adjusted
<J23k87> that .9° stepper should limit the vfa (vertical fine artifacts)
<J23k87> did they have input shaping or is that only for the XL
<teepee> yep, that's what it said in a video I watched yesterday about mk4 vs. bamboolabs
<teepee> yes, that tree was printed with input shaping 0.4mm nozzle < 2h
<teepee> let me slice that with my old profile in cura
<teepee> 4h
<teepee> not sure if the input shaping does much for this case though
<teepee> next up, a star shaped gift box :)
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dalias has joined #openscad
LordOfBikes has joined #openscad
J23k87 has quit [Quit: Client closed]
J23k87 has joined #openscad
mmu_man has quit [Ping timeout: 256 seconds]
mmu_man has joined #openscad
othx has quit [Remote host closed the connection]
othx has joined #openscad
<teepee> ok, 2 days covered, 22 to go :-)
<InPhase> Oh, started calendaring?
<InPhase> Bit late, but we can catch up. :)
<InPhase> Did you Mastadon-out or whatever their word is and mailing list announce to try to gather some extras?
<teepee> pff, how is that late, there's even more November days left :D
<teepee> no, not yet
<teepee> I guess that would mean at least some sort of "go, go, go, we will make it" :)
<teepee> like from the current members of the santas-helpers team
<teepee> I do have a multi-part idea that might work still - I don't think I can finish any of the other two that are waiting for years already
<teepee> I saw a model like that not long ago, which should be realtively easy to do
<teepee> one of those old style lamps with 4 different images
dalias has quit [Ping timeout: 255 seconds]
<gbruno> [github] gsohler synchronize pull request #4842 (Indexed2 squash) https://github.com/openscad/openscad/pull/4842
<teepee> not as cool as that one, but openclipart hopefully has some images that should work
dalias has joined #openscad
misterfish has quit [Ping timeout: 255 seconds]
greenbigfrog has quit [Ping timeout: 256 seconds]
<teepee> hmm, the lid is not clipping in as well as I want, but otherwise it's fine https://imgur.com/a/2A93V5e
greenbigfrog has joined #openscad
misterfish has joined #openscad
zauberfisch has joined #openscad
dalias has quit [Ping timeout: 256 seconds]
dalias has joined #openscad
<InPhase> teepee: Run a gluestick around the outer edge of the insert lip if the lid, then let it dry fully. Then observe it magically stay in place better. :)
<InPhase> It's like printing a thin grippy layer. ;)
<InPhase> File under 75% chance of success.
<InPhase> Washable if it fails to help.
mmu_man has quit [Ping timeout: 256 seconds]
misterfish has quit [Ping timeout: 264 seconds]
mmu_man has joined #openscad
kintel has joined #openscad
<teepee> haha, magic closed box
<teepee> printed a new lid meanwhile, perfect now :P
<teepee> 3 days
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zauberfisch has quit []