#hazzy Logs

Aug 24 2020

04:53 AM Joco[m]: hazzy: pretty to crash it using a large file. Takes a long time for the syntax highlighting to work its way through a 2MB plus file.
04:53 AM Joco[m]: * hazzy: pretty easy to crash it using a large file. Takes a long time for the syntax highlighting to work its way through a 2MB plus file.
04:57 AM Joco[m]: in your tinkerings did you try explicitly to delink the doc that the syntax highlighter is refering to? And/Or did you try creating a new syntax highlighter instance?
04:57 AM Joco[m]: * in your tinkerings did you try explicitly to delink the doc that the syntax highlighter is refering to? And/Or did you try creating a new syntax highlighter instance? Like you do when using a new font
05:17 AM JT-Cave: morning
05:30 AM hazzy-m: Morning
05:31 AM hazzy-m: Joco: yes I did, no luck
06:04 AM JT-Cave: hmm I need to create a hex sequence like 0x010101 but hex() strips the leading 0's off in python
06:05 AM hazzy-m: are you trying to get it as a string?
06:06 AM JT-Cave: no a hex
06:06 AM hazzy-m: I've run into that before, I think I ended up using on the for format specifiers and got what I wanted
06:06 AM JT-Cave: s[0] = 0x010101
06:07 AM JT-Cave: that sets each led to level 1
06:07 AM JT-Cave: there are three leds per section on the dotstar
06:10 AM JT-Cave: >>> x = hex(1)[2:].zfill(2)
06:10 AM JT-Cave: >>> type(x)
06:10 AM JT-Cave: <class 'str'>
06:10 AM JT-Cave: >>> type(0x01)
06:10 AM JT-Cave: <class 'int'>
06:11 AM JT-Cave: x = hex(1)[2:].zfill(2) + hex(1)[2:].zfill(2) + hex(1)[2:].zfill(2)
06:11 AM JT-Cave: s[0] = int(x)
06:12 AM JT-Cave: that works but seems a bit kludgy
06:39 AM JT-Cave: damn that does not work it converts it to 1
02:48 PM TurBoss: hello
03:08 PM satiowadahc_work: I'm going to assume the qtvcp in the linuxcnc repository isn't being updated to python3 at the moment?
03:32 PM TurBoss: hello , there is a branch of QtPyVCP with python 3 support its a bit old but if runs
03:32 PM TurBoss: because we only support one version of python 2 or 3
03:32 PM TurBoss: wi will wait to a release from linuxcnc
03:34 PM satiowadahc_work: Python2 support is gone no?
03:34 PM TurBoss: we still need python2
03:34 PM TurBoss: python 2 is end of life
03:34 PM TurBoss: yep dead
03:35 PM TurBoss: didn't took to much to port 2 to 3
03:35 PM TurBoss: on qtpyvcp
03:36 PM TurBoss: linuxcn had a brutal amount of work done by Rene Hopf (Gitter)
03:36 PM satiowadahc_work: Is there much qifference between qtvcp and qtpyvcp?
03:36 PM satiowadahc_work: Yea I've been following some of there work these last few days, currently trying to get my own repository up to date
03:37 PM TurBoss: while Qtvcp uses Glib signals wich came from gtk (gtk is not a requirement anymore)
03:37 PM TurBoss: we use Qt signals
03:39 PM satiowadahc_work: Ahh that makes sense!
03:39 PM TurBoss: i don't know qtvcp to know more diferences
03:40 PM satiowadahc_work: I'm going to be diving into it off and on for a few weeks, been granted 120 hours to dedicate to open source right now
03:41 PM TurBoss: do you plan to create your own UI
03:41 PM TurBoss: ??
03:41 PM satiowadahc_work: Already have :)
03:42 PM satiowadahc_work: It was based on GTK ported to QT and now its kinda a plate of spaghetti
03:42 PM TurBoss: https://github.com/kcjengr/ ;)
03:42 PM satiowadahc_work: I feel like that should be in the topic of this channel... I can never seem to find that link
03:44 PM TurBoss: hazzy: started working on "Hazzy" ui gtk2 back in 2015 the we moved to Gtk3 and then started over with Wt
03:44 PM TurBoss: * hazzy: started working on "Hazzy" ui gtk2 back in 2015 the we moved to Gtk3 and then started over with Qt
03:44 PM satiowadahc_work: Ya'll need to stick to a plan :P
03:45 PM TurBoss: 😇
03:45 PM TurBoss: http://www.qtpyvcp.com/showcase/mill_vcps.html
03:46 PM * TurBoss uploaded an image: Screenshot from 2020-08-22 14-27-52.png (81KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/RkCEmbiYVefwwwpMzQjJYewy/Screenshot from 2020-08-22 14-27-52.png >
03:47 PM satiowadahc_work: Maybe if I get clearance from higher ups I'll share screen shots...
03:47 PM satiowadahc_work: I dislike proprietary but they pay me so ill deal with it
03:47 PM TurBoss: no problem
03:50 PM satiowadahc_work: So many broken things in qtvcp right now...
03:54 PM satiowadahc_work: I wonder if theres any benefit in even updating qtvcp to 3?
04:03 PM satiowadahc_work: Why did they pout a lib module in a lib module???
04:03 PM satiowadahc_work: *puy
04:03 PM satiowadahc_work: *put
04:05 PM TurBoss: what where?
04:05 PM TurBoss: :)
04:05 PM satiowadahc_work: https://github.com/LinuxCNC/linuxcnc/tree/master/lib/python/qtvcp/lib
04:05 PM satiowadahc_work: its looking at the first lib not the second in python3
06:01 PM hazzy-m: satiowadahc_work: you should switch to QtPyVCP :)
06:20 PM JT-Cave: I second that, qtpyvcp is much better and well thought out
06:46 PM Joco[m]: hazzy: i'm tinkering with the seg fault issue. I have found one reference on line to what smells like what is happening. If true its a QT bug
06:49 PM TurBoss: Joco: what? you got segfault?
06:50 PM Joco[m]: the plan text control syntax highlighter widget will seg fault if you load a large doc and the load something else before it completes its syntax highlighting. hazzy found it.
06:50 PM TurBoss: :O
06:50 PM Joco[m]: On a 200k line file it take 10+ minutes to process. Possibly longer, I gave up waiting after 10-15 mins
06:51 PM Joco[m]: * On a 200k line file it take 10+ minutes to process. Longer, I gave up waiting after 10-15 mins
06:51 PM TurBoss: hmmm
06:51 PM Joco[m]: python just an't that fast
06:51 PM Joco[m]: need cython
06:51 PM Joco[m]: ;-)
06:52 PM TurBoss: maybe cythonize the whole project no but some parts?
06:53 PM TurBoss: it compiles and runs but some issues with relative paths "__file__"
06:54 PM Joco[m]: anyway the reference I found was to the destructor of the syntaxhighlighter trying to reference/write back to the textdocument it was linked to. So if that doc is gone ... bam
06:55 PM Joco[m]: This would explain why it seg faults in a QT lin
06:55 PM Joco[m]: * This would explain why it seg faults in a QT lib
07:04 PM TurBoss: https://gist.github.com/ctokheim/6c34dc1d672afca0676a cython tricks
07:04 PM roguish[m]: hey you guys. what do you think is a really good sub-$500 3D printer? one worth buying........
07:14 PM TurBoss: i like the cetus 3d
07:14 PM Joco[m]: hazzy: I have isolated the problem to the Stop/Start methods on the syntax highlighter. If I comment them out of the setPlainText method I can load a ~3MB file (200k lines) then load a nice small one. It's slow to sort out whatever it is doing internally but after 15-30secs the new file loads. No seg fault.
07:14 PM TurBoss: I don't have one
07:14 PM * TurBoss uploaded an image: image.png (127KiB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/dxHnuPYUiXxFwpcUDQkfGrCO/image.png >
07:22 PM TurBoss: roguish: there is also the snapamaker
07:28 PM TurBoss: * roguish: there is also the snapmaker
08:41 PM TurBoss: https://github.com/Technologicat/setup-template-cython/blob/master/setup.py
10:26 PM hilo90mhz[m]1: roguish: I would look at the Creality Ender 5 Pro in that price range
10:26 PM hilo90mhz[m]1: You want a more rigid machine than the single arm type like the Cetus in my opinion. I print in high speeds and it does matter..
10:26 PM hilo90mhz[m]1: I don't have either printer though
10:28 PM * hilo90mhz[m]1 uploaded an image: 61Cc0N9RfyL._SL1001_.jpg (72KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/WGfFsqlCzxbPfAhqdsqThRLv/61Cc0N9RfyL._SL1001_.jpg >
10:28 PM hilo90mhz[m]1: I have some heavily modified Makergear M2 printers - wouldn't recommend them these days but ive had mine for almost 10 years now
10:38 PM Joco[m]: thats essentially what I have the parts to built
10:38 PM Joco[m]: a hypercube style
10:39 PM Joco[m]: got all the ali extrusion and stuff
10:39 PM Joco[m]: a mate has already built his using the same patrts
10:39 PM Joco[m]: will be a big update from my existing reprap prusa
10:41 PM hilo90mhz[m]1: yeah the cube type are nice - I have some parts and want to build a large very rigid one someday
10:41 PM hilo90mhz[m]1: but if you do need something turnkey its hard / impossible to beat the $400 price for the whole printer those sell for
10:41 PM hilo90mhz[m]1: with magnetic bed which is key these days
10:52 PM Joco[m]: what's this magentic bed>
10:52 PM Joco[m]: ?
10:54 PM Joco[m]: found it. Nice. Spring steel, pre applied PEI covers. Cool.,