fling has quit [Remote host closed the connection]
fling has joined #openscad
snaked has joined #openscad
epony has joined #openscad
linext has joined #openscad
lostapathy has quit [Quit: WeeChat 3.5]
lostapathy has joined #openscad
linext__ has quit [Ping timeout: 276 seconds]
linext_ has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
linext has quit [Ping timeout: 260 seconds]
LordOfBikes has quit [Ping timeout: 268 seconds]
mmu_man has quit [Remote host closed the connection]
mmu_man has joined #openscad
LordOfBikes has joined #openscad
snaked has quit [Ping timeout: 240 seconds]
cool has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
snaked has joined #openscad
cool has quit [Quit: cool]
J23k70 has joined #openscad
J23k63 has quit [Ping timeout: 250 seconds]
L29Ah has quit [Ping timeout: 260 seconds]
kintel has joined #openscad
<kintel>
pca006132 In terms of building a new/modified parser/ast for OpenSCAD: It would be interesting to discuss whether we can package that as a library and license/relicense it under a more no-strings-attached license, mostly to promote interop with surrounding non-GPL ecosystems.
qeed has quit [Quit: qeed]
qeed has joined #openscad
cool has joined #openscad
pca006132 has quit [Remote host closed the connection]
cool has quit [Remote host closed the connection]
pca006132 has joined #openscad
<pca006132>
kintel: I am not entirely sure if it is possible (if we apply GPL in a strict way), as I have read the source code of openscad and the related code I write is counted as derived work... (and the parser grammar is very similar)
<pca006132>
but I am fine for re-licensing the code I write
<kintel>
pca006132 yeah, we'd have to first relicense the openscad parser etc. - it would require some legwork but it might be possible. Anyway, no need to jump at it, but good to open the discussion
<pca006132>
sure
<teepee>
is there a lexer with utf-8 support by any chance?
<pca006132>
btw the parser is far from complete, I am just attempting to handle simple expressions for now
<pca006132>
the lexer is in much better shape, just missing use
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pca006132>
the bison c++ API is surprisingly good, albeit it is lacking detailed documentation and more complicated examples
<pca006132>
btw, the existing parser code is quite confusing: things like echo are valid module identifiers, but this is merely a hack to let users use them in places that expects module instantiation...
<guso78k>
it should display two cubes, but displays only one. Unlike bare openscad, in pythonscad sub nodes can be reused and in this case both cubes are translated accidently. Any Clue where to modify the copy instead of the the original ?
misterfish has joined #openscad
<guso78k>
this only applies for preview mode. it does well in render.
<pca006132>
teepee: my lexer now supports utf-8 properly, with correct position handling and identifier classification
<pca006132>
at least for languages from left to right, I have no idea how columns work with BIDI and row works with Thai
<pca006132>
(I just count the number of graphemes for the column number)
<guso78k>
pca006132, its very compact. flex is a nice language!
<guso78k>
pca006132, look at this: https://imgur.com/a/eQUTqFC this is similar to relativity package in openscad but with a python class based approach, where you can literally stick objects together(any transformation possible with numpy matrices)
<pca006132>
will the position attributes be updated when you apply transform to the (outer) object?
<pca006132>
this can simplify a lot of calculation
<guso78k>
pca006132 not yet, but i can work on it. once placed its stick for the moment
arebil has quit [Quit: Bye]
<guso78k>
every object still has the absolute transformation available, so objects can easily be re-transformed
<guso78k>
pca006132 probably i got your question wrong. the very top stick is bent by 60 degree, this is the *implicit* result from adding 30 and 30. so you dont have to worry about this. you extend your model at various places as you add code without having to care about translations. np.matmul does the trick
<pca006132>
ok nice
<guso78k>
now adding some python decorators to make usage more nice
arebil has joined #openscad
RichardPotthoff has quit [Ping timeout: 256 seconds]
erectus has quit [Quit: erectus]
drkow has joined #openscad
drfff has quit [Ping timeout: 260 seconds]
cart_ has joined #openscad
misterfish has quit [Ping timeout: 245 seconds]
ril3y has quit [Quit: Client closed]
nickmcski has quit [Quit: Client closed]
L29Ah has joined #openscad
misterfish has joined #openscad
guso78k has quit [Quit: Client closed]
J23k70 has quit [Quit: Client closed]
J23k70 has joined #openscad
epony has quit [Remote host closed the connection]
guso78k has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
guso78k has quit [Ping timeout: 250 seconds]
misterfish has quit [Ping timeout: 268 seconds]
misterfish has joined #openscad
snaked has joined #openscad
teepee has quit [Remote host closed the connection]
<teepee>
hmm, so that's while running cmake. I think that means it should be fine with getting it installed in the base image as normal linux program
<teepee>
e.g. no need for having a cross-compiled version
<teepee>
package zstd
<guso78k>
yes, but the red color branch-build check for MXE32 and MXE64 is offensive even though its not really needed to get it clean(as you state). this is why i am interested to have it fixed
<guso78k>
in my place creating and using MXE64 bit installers works fine
<teepee>
it's designed that way, so it should not be too complicated, but DIY was the default setting and we just missed clicking to the pre-built one
<teepee>
but I guess it's good in the end as this is basically the test run to see if that may become the new default laptop for us
fling has quit [Remote host closed the connection]
fling has joined #openscad
<InPhase>
teepee: The framework laptops look pricey, but I do really appreciate the general design and clear ease of obtaining replacement and repair parts.
<InPhase>
teepee: Please let me know how it seems to you once you get it up and going.
<InPhase>
teepee: Like, does it end up seeming super sturdy? Or do you think some of these parts could snap off from flimsy plastic tabs and stuff under physical use?
guso78k has quit [Quit: Client closed]
<InPhase>
Obviously the raw components offered seem pretty decent, so I'm more concerned about the mechanical aspects of that system.
<InPhase>
(Unless there is some surprise on the components.)
<InPhase>
teepee: Also I have questions about the thermals. The arrangement there for thermal management doesn't quite look optimized, so I'm curious how it holds up under a sustained heavy load.
<teepee>
yep, I heard reasonable good things, so that's now a first look at something else. the Tuxedo we have so far are not bad and I like the Linux theme, but there's always small trouble over time
<pca006132>
my wish is a laptop with great battery life, sadly the only choice now is a macbook...
<InPhase>
teepee: Yeah. I actually have an outstanding service request with Tuxedo about getting a keyboard replacement... Basically I'm just sitting around waiting for them to get around to processing my service request from quite some time ago. And that's really not an acceptable thing to have happen on a primary computer. I'm still able to use it, but if I weren't, that would be a serious problem.
<teepee>
not going to touch those anymore, I did have a macbook pro for quite a while - about 6 years I think?
<pca006132>
progress update: the parser I wrote can now parse everything in openscad except for loop expressions, conditional expression and list related stuff (comprehension blablabla), and it is under 1k lines in total including the AST definitions
<pca006132>
I hope that I can get something under 3k lines in the end...
<teepee>
InPhase: yeah, for us it's mostly tiny stuff like a base of the case breaking off small bits that go into the fan and it's difficult to get that repaired
<InPhase>
teepee: If they could hire enough staff to actually ANSWER emails then Tuxedo would be great. I gave them a pass when they had trouble with this under Covid, but seriously, it's Dec. 2023 now... If you can't hire enough service personnel right now, then you're not trying.
<teepee>
pca006132: what you mentioned earlier the strange echo handling and stuff, I think that comes from when the listcomprehensions were added
<InPhase>
teepee: Ah. I got a metal frame one from Tuxedo, so it has been super sturdy mechanically. I had a battery issue, that they replaced for "free", except they took so long I had to buy a third party one. So now I have a free spare battery for this once theirs finally arrived.
<teepee>
if therer's an option to reduce those, like not having them as keywords, would be neat
<InPhase>
teepee: So I give them credit for replacing it free, but take away points for a response too slow to actually work.
<InPhase>
teepee: And now my S key keeps falling off. This is an understandable problem. It's a 2.5 year old laptop, and sometimes keyboards break on that timescale. I dont fault them for that part. But I should be able to just place an order for a new keyboard and get it sent.
<teepee>
InPhase: not *that* easy in germany, my car shop did complain too when handing out almost 2 month waiting times
<teepee>
InPhase: exactly and the hope is that with more people having that notebook series we can just have a keyboard and battery sitting on the shelf to swap out when needed
<InPhase>
teepee: Frankly, I don't care where their service people work! If they can't hire people in Germany for this, then they should hire elsewhere. These people just need to manage electronic service requests competently.
<InPhase>
teepee: If they're paying too little, then they should fix that. Their stuff isn't so cheap that they should be unable to afford labor.
<pca006132>
I have no idea why these customer service are so slow when the parts are in stock
<InPhase>
It's the single thing that is nudging me to shop elsewhere. I was even willing to overlook the slow shipping from Germany. :)
<pca006132>
it is not like they are making it from scratch in the factory...
<InPhase>
pca006132: Process management is apparently its own skill.
<InPhase>
pca006132: And some people just don't have it.
<InPhase>
I suppose Tuxedo is a small company, and whoever is managing it just doesn't understand the problem. I'd happily explain it, but they probably wouldn't even read the message for 6 months...
<pca006132>
just hire someone that is competent...
<teepee>
that's a hard skill to get, as obvious by many companies ;-)
<pca006132>
I feel that process management in larger companies can be more painful due to the additional overhead from their large size
<InPhase>
But they've surely heard it before anyway. I see such comments online about them. Perhaps they've just excused it away, and don't see that it's the one thing hindering their growth. They could be making so much more money.
<teepee>
yep, they would be able to keep us as long term customer, not a huge amount of people but still maybe 15 notebooks going around that need replacements from time to time
<InPhase>
teepee: But anyway, give me a detailed framework laptop impression/review when you have one. I'm a little wary of only getting 4 ports to work with, but otherwise excited about the concept.
<teepee>
wait for the 16 inch one, IIRC that has 6 ports
<InPhase>
Oh. That'd do it.
<teepee>
right now I have 3 ports, so I'm good :)
<teepee>
and I can even have a USB-A port
<InPhase>
I was thinking about the 13, figuring it would be nice to have one smaller laptop. But if the 16 has an acceptable mass, 6 ports could be a deciding factor.
<teepee>
I'm used to that with my XPS-13, but it's really slowly dying now. after 6 years!
<InPhase>
I require one ethernet and one hdmi, and I require at least one USB-C and at least one USB-A, which then gives a frustratingly small flexibility if I only have one of each type of USB port.
<InPhase>
Two of each USB port type would be spectacular.
<pca006132>
6 years is quite impressive, did you replace any parts on the XPS-13?
<teepee>
yes, battery puffed up (now again), keyboard had a key broken and something on the display, not sure what it was
<teepee>
so we had 3 rounds of not-quite-next-day-repair for the warranty
<InPhase>
I crossed off the XPS-13 as an option in a prior search when I saw how badly they were throttling. Although I think that was the late 2020 version of it.
<teepee>
but I had never a day where I had to give it away
<teepee>
yeah, I have 2nd generation I think and the thermals are quite bad
<teepee>
I usually run it manually throttled as for most stuff that's just fine
<teepee>
unlimited it goes to 100° cpu thermal limit which is not a good thing
<InPhase>
I've learned to highly prioritize trying to figure out the thermal performance before purchase. But that's a tricky thing to find out without having it. You have to get lucky on a youtube video review.
<InPhase>
I wish it would become standard to report that as a spec.
<pca006132>
my current laptop is a XPS-15, the thermo is not great but usable
<InPhase>
OpenSCAD has not been straining thermals for me outside of compilation, but llama.cpp sure does. :)
<teepee>
minus the thermal issues, the xps-13 is quite a nice machine with it's 4k display
<teepee>
:)
<pca006132>
InPhase: try more minkowski huge huge $fn!
<pca006132>
*with huge*
<InPhase>
:)
<pca006132>
manifold will happily eat all your cores
<InPhase>
pca006132: Sure thing! As soon as manifold stops breaking on the cache issue. ;) That one will keep me from using it as a daily driver.
<pca006132>
the scaling is still pretty limited though, because the mesh simplification is sequential
<InPhase>
Oh right. I was going to add my design to that issue last night, but got distracted by several other things. I'll try again tonight.
<pca006132>
I think we need someone to look into the cache issue, I am busy working on sscad right now so probably no time for that (and I am not familiar with the cache part anyway)
<pca006132>
the goal is to be able to run some simple scripts before next year
<J23k70>
did you print the sides diagonal or how do they fit on the bed?
<J23k70>
if you had enough clearance - it should work if you don't spill glue.. maybe use a oil q-tip to protect the parts that shouldn't glue together
splud has joined #openscad
<teepee>
yes, sized to just about fit the bed diagonal
<teepee>
the separate SVGs created by MichaelFrey who did contribute to openscad earlier. Small world :)
<teepee>
although I ended up using the combined SVG anyway as the separation is just hiding the other one outside the page area and OpenSCAD ignores that when importing
<pca006132>
wonder how different it must be to not be considered derivative work by GPL, feel that relicensing the parser in current openscad is going to take a long time
<teepee>
maybe not too many people did bigger changes
<teepee>
that said, I'm not sure how such a description of an API is handled copyright wise
<pca006132>
I remember people are saying things like
<pca006132>
if you read the source code and implement something similar in functionality, it is up to the court to decide
<pca006132>
something like that
<pca006132>
because you cannot erase your memory
<teepee>
hmm, more people involved that I would have thought
<teepee>
yeah, plus depending on legislation there's some extra burden to show a minimum of effort, e.g. germany has that
<teepee>
so things like changing the comma behavior would probably not introduce copyright in german court
<pca006132>
feel that it is extra ambiguous for things like a parser, you only have so many (sane) ways to write them
<teepee>
yes, exactly
<teepee>
and there's probably not any good way to get a safe answer
<pca006132>
imagine telling the judge what CFG means and the usual ways of writing a parser
<pca006132>
probably not easy
<teepee>
uh, how do I get a discord notification when not logged in for maybe a couple of month
<teepee>
anyway, need to get some groceries and find out if that package is actually the new notebook :)
pca006132 has quit [Quit: pca006132]
J23k70 has quit [Quit: Client closed]
J23k70 has joined #openscad
pca006132 has joined #openscad
<pca006132>
wow I am amazed, conan and vcpkg are marketing as nice c/c++ package manager but they don't even support bison properly...
arebil has quit [Quit: arebil]
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
erectus has quit [Read error: Connection reset by peer]
erectus has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
<teepee>
ok, notebook confirmed, so tomorrow is going to be build day ;-)