#hazzy Logs
Oct 01 2019
#hazzy Calendar
01:47 AM Lcvette: Lcvette fell asleep at the computer
01:48 AM Lcvette: Tom_L (@freenode_Tom_L:matrix.org): what's that link?
01:48 AM Lcvette: Turbos if we add the offset parameter numbers manually to the var file, won't it populate the offsets?
01:49 AM Lcvette: If that can read the car file directly that could work no?
02:02 AM TurBoss: I'm taking a look to the tooltable
02:02 AM TurBoss: it does the same
05:06 AM jthornton: morning
05:08 AM TurBoss: HI
05:27 AM hazzy-m: morning
05:27 AM Tom_L: hey
05:44 AM TurBoss: Tom_L: I started working on the offset table with the code you pasted
05:44 AM TurBoss: its very usefull
06:13 AM hazzy-m: Hurray!
06:17 AM TurBoss: I'll just read the var file
06:40 AM * TurBoss uploaded an image: Captura de pantalla de 2019-10-01 13-34-30.png (113KB) < https://matrix.org/_matrix/media/v1/download/matrix.org/mtcLPOKBxYbfJGvYzOwDUICk >
06:41 AM * TurBoss uploaded an image: Captura de pantalla de 2019-10-01 13-35-09.png (7KB) < https://matrix.org/_matrix/media/v1/download/matrix.org/PHnJqtFioJKTTtCxMvlBIryf >
06:42 AM * TurBoss is burning its cpu
06:55 AM hazzy-m: TurBoss: I have a branch with a plug-in that watches and reads the car file, but it never gets written too
06:56 AM hazzy-m: I think the branch is INTERP_PARAM_PLUGIN
06:59 AM TurBoss: ok
06:59 AM TurBoss: ty
09:15 AM Lcvette: \o/
09:15 AM Lcvette: thats great news turboss!
09:16 AM TurBoss: Lcvette: can you test a thing? on the sim?
09:22 AM Lcvette: sure
09:25 AM Lcvette: TurBoss: i don't see anything to checkout
09:26 AM TurBoss: no need
09:27 AM Lcvette: i have atc_smi open
09:27 AM TurBoss: can you change offsets there?
09:28 AM Lcvette: i don't know what you are asking me
09:28 AM Lcvette: with the offset buttons?
09:29 AM Lcvette: G54 G55
09:29 AM Lcvette: yes
09:29 AM TurBoss: doesn't matter
09:29 AM Lcvette: they are all the same as machine coordintate
09:29 AM Lcvette: do you want me to save some different coordinates to them?
09:29 AM TurBoss: yes
09:29 AM TurBoss: but before
09:30 AM TurBoss: open the var file in some editor
09:30 AM TurBoss: it will show us if they are writen in realtime
09:30 AM Lcvette: vmc.var
09:30 AM Lcvette: opened
09:31 AM TurBoss: ok
09:31 AM TurBoss: now change the value of something
09:31 AM TurBoss: a offset
09:32 AM Lcvette: 5220 5221 and 5222 all changed to the new value when i zerod the axis
09:33 AM TurBoss: the editor showed the changes?
09:34 AM * Lcvette uploaded an image: Screenshot_2019-10-01_10-28-12.png (41KB) < https://matrix.org/_matrix/media/v1/download/matrix.org/hrVGfYLrhHhEFfTLvczOfWWm >
09:34 AM TurBoss: ok
09:35 AM Lcvette: i did G0 X1 Y1 Z-1 in G54
09:35 AM Lcvette: then i did G0 X2 Y2 Z-2 in G55
09:35 AM Lcvette: both showed up in the opened var file
09:36 AM TurBoss: nice
09:37 AM TurBoss: thats
09:37 AM TurBoss: thanks!
09:37 AM Lcvette: just did the same for G56 G0 X3 Y3 Z-3 then seroed the axis and it wrote to the var file
09:37 AM TurBoss: I was thinking that the var file will be write on ly on close or something ....
09:37 AM Lcvette: \o/
09:38 AM Lcvette: maybe it writes offsets when you zero the axis?
09:39 AM Lcvette: can you use the same thing you used on the atc to print the numbers to the gui?
09:40 AM Lcvette: in the table
09:40 AM TurBoss: but that will require to issue some mcode?
09:40 AM Lcvette: ah
09:41 AM Lcvette: could it be tied to the axis zero button so it writes anytime the axis is zeroed?
09:42 AM Lcvette: i think it changes the var file but doesn't save it until shutdown
09:42 AM Lcvette: if i read correctly in my research last night
09:43 AM TurBoss: if you saw them on the editor they are saved
09:43 AM Lcvette: so it will change the var file but if an uncontrolled shutdown of linuxcnc occurs the paameters will be lost
09:43 AM Lcvette: ?
09:43 AM Lcvette: ok
10:07 AM Lcvette: still there?
10:07 AM Lcvette: have an idea?
10:10 AM TurBoss: tellme
10:10 AM Lcvette: im asking
10:10 AM TurBoss: ah
10:11 AM TurBoss: yes I'll merge what hazzy did and try
10:12 AM Lcvette: cool
10:20 AM TurBoss: but I also did things
10:21 AM TurBoss: mostly the same
11:18 AM Lcvette: you have a idea on how to do the offsets table then?
11:38 AM TurBoss: yes
11:38 AM TurBoss: like the tool table
11:38 AM TurBoss: a QTableView widget
11:39 AM TurBoss: with options to enable columns etc...
11:57 AM tommylight: evening
11:58 AM TurBoss: hello
11:58 AM tommylight: sup Boss ?
12:08 PM Lcvette: nice turboss that will be slick!!
12:42 PM Tom_L: makin progress?
12:59 PM TurBoss: yes
01:00 PM TurBoss: now away but making progress
01:00 PM TurBoss: managed to read on boot
01:01 PM Tom_L: nice
01:01 PM Lcvette: thats perfect!
01:02 PM Lcvette: if you can read stored offsets on boot, then you can read current work offsets when they are set? could use the available gcode to set the newest offset to keep everything current in the table right?
01:21 PM Tom_L: that file also read the current offsets from memory
01:21 PM Tom_L: the inactive ones must be read from the .var file
01:25 PM Tom_L: linuxcnc does not give
01:25 PM Tom_L: # access to all offsets thru NML, only current ones
01:42 PM TurBoss: i tried to add that feature
01:42 PM TurBoss: to the linuxcnc
04:07 PM * TurBoss sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/zndzzxnyvjIzKPguCkvWnaTm >
04:08 PM Lcvette: you did it?
04:14 PM TurBoss: just got the data on screen
04:14 PM Lcvette: thats awesome!
04:39 PM Tom_L: TurBoss, is that direct from the array?
04:40 PM TurBoss: from the params file
04:40 PM Tom_L: do you know what each element represents?
04:41 PM Tom_L: that first one must only have a Z offset?
04:42 PM TurBoss: from 1 to 9 they are the offset index and the list in each row is the offset of each axis
04:43 PM Tom_L: acording to the gcode docs it is one off
04:43 PM Tom_L: should be 10
04:43 PM Lcvette: \o/
04:43 PM Tom_L: x,y,z,a,b,c,u,vw,r
04:43 PM Tom_L: x,y,z,a,b,c,u,v,w,r
04:44 PM Lcvette: r is rotation
04:44 PM Tom_L: pretty sure that's standard RS274 format
04:45 PM Tom_L: it should include them all, who knows how ppl will use it
04:45 PM * TurBoss uploaded an image: image.png (28KB) < https://matrix.org/_matrix/media/v1/download/matrix.org/woigyjHLSGplpHWMGThEmRPt >
04:45 PM Tom_L: http://linuxcnc.org/docs/2.8/html/gcode/g-code.html#gcode:g54-g59.3
04:46 PM Tom_L: as long as you're not using them i suppose that's fine as long as they get written back to the file in the right location
04:47 PM Tom_L: looks like you're making good progress though
04:47 PM Lcvette: TurBoss: the r is a good idea
04:48 PM Lcvette: since probing can auto set a rotation
04:48 PM Tom_L: his example code i posted only uses 9 though
04:48 PM Tom_L: not sure why
04:49 PM TurBoss: rotations is stored in another place
04:49 PM TurBoss: not the same table but similar
04:49 PM TurBoss: but R is for each axis?
04:49 PM Tom_L: just saying it has a place holder in the .var file
04:49 PM Tom_L: yes
04:49 PM Tom_L: check the doc link
04:50 PM TurBoss: oh I was wrong
04:50 PM TurBoss: R is there
04:50 PM TurBoss: is just read 1 step mre
04:50 PM TurBoss: * is just read 1 step more
04:52 PM Tom_L: 5230 0.000000
04:52 PM Tom_L: 5230 0.000000
04:52 PM Tom_L: the one for x
04:52 PM Tom_L: err G64
04:52 PM Tom_L: sry
04:52 PM Tom_L: 54
04:52 PM Tom_L: dammit
04:53 PM Tom_L: must be a tab between them
05:01 PM Tom_L: yep, just checked... 0x09
05:02 PM Tom_L: with 0x0A at the end (LF)
05:04 PM Tom_L: i suppose when you save them to file you would only be saving the current active offset
05:04 PM Tom_L: i think that's how axis works but not positive
05:18 PM Not-5e4d: [02qtpyvcp] 07TurBoss created branch 03offset_table - 13https://git.io/fhCou
05:18 PM Not-5e4d: [02qtpyvcp] 07TurBoss pushed 035 commits to 03offset_table [+2/-0/±8] 13https://git.io/Jen5m
05:18 PM Not-5e4d: [02qtpyvcp] 07TurBoss 034e1a8d5 - offset table
05:18 PM Not-5e4d: [02qtpyvcp] 07TurBoss 03c2561a5 - enable vertical header
05:18 PM Not-5e4d: [02qtpyvcp] 07TurBoss 03fcf817e - more refactor
05:18 PM Not-5e4d: [02qtpyvcp] ... and 2 more commits.
05:20 PM travis-ci: kcjengr/qtpyvcp#1097 (offset_table - 3b811fd : TurBoss): The build passed.
05:20 PM travis-ci: Diff view : https://github.com/kcjengr/qtpyvcp/compare/4e1a8d55c5a1^...3b811fdd02f6
05:20 PM travis-ci: Build details : https://travis-ci.org/kcjengr/qtpyvcp/builds/592277021
05:27 PM Tom_L: does that write out the entire file?
05:28 PM Tom_L: or just read mode for now?
05:36 PM Lcvette: Turboss you got it working?
05:40 PM Tom_L: note that the .var file starts at 5161 and ends at 5390 but the coordinate system starts at 5221 for G54 and ends with 5390. other system data is stored in the .var file along with the fixture offsets
05:41 PM Tom_L: iirc some are set aside for user vars as well
05:43 PM TurBoss: im having a break
05:43 PM Tom_L: also the value at 5220 (1-9) determines what offset the machine will start up with and defaults to 1.00000
05:43 PM Tom_L: can't say i blame ya :)
05:45 PM Tom_L: i wish i knew python
05:52 PM Tom_L: i see the if elif that search for the right ones now...
05:53 PM JT[m]: what are you guys doing?
05:53 PM Lcvette: offset table
05:53 PM Tom_L: TurBoss is working on reading/writing the .var file
05:53 PM Lcvette: JT \o/
05:55 PM JT[m]: you gotta be careful when you do that because there is no guarantee when linuxcnc writes to it
05:55 PM JT[m]: you could be reading stale data if you read it after startup
05:56 PM Lcvette: the offsets should be fresh no?
05:56 PM Lcvette: they are persisten until changed
05:56 PM Lcvette: or at least that is what i have seen in thevar file after start up
05:57 PM JT[m]: but the var file is not updated when you change while running, and yes at startup you should be ok until you change something
05:57 PM JT[m]: after you set any variable in the var file it's no telling when linuxcnc will actuall write that to the file
05:58 PM Lcvette: for the current work offset i think you can use an mdi command to set it and he is working on using that same command to update the var
05:58 PM Lcvette: JT we have been seeing real time var updates when setting work offset
05:58 PM Lcvette: could it be something new in 2.9?
05:58 PM JT[m]: no been like that forever
05:58 PM Tom_L: JT[m], this is what i pointed him to as a guide: https://github.com/LinuxCNC/linuxcnc/blob/master/lib/python/gladevcp/offsetpage_widget.py
05:58 PM Tom_L: for glade i think
05:59 PM Lcvette: turboss and i tested setting offsets with the var file open and so far 100% of the time the var file update as soon as the axis was zeroed
06:00 PM JT[m]: but it's not guaranteed to do that everytime... and that was a big topic in the past
06:00 PM Lcvette: do you mean it saves randomly or updates randomly?
06:00 PM Lcvette: oh
06:00 PM Lcvette: hmm
06:00 PM Lcvette: they can't fix it?
06:01 PM Tom_L: anybody is free to fix it
06:01 PM Lcvette: seems like it would be beneficial
06:01 PM JT[m]: who?
06:01 PM Lcvette: the people who made it
06:01 PM JT[m]: exactly
06:01 PM Lcvette: or who update it
06:01 PM JT[m]: NIST?
06:01 PM Tom_L: maybe they got tired
06:02 PM Lcvette: where do the new releases come from?
06:02 PM Tom_L: volunteers
06:02 PM JT[m]: about the only person that might be able to understand it is Dewey
06:02 PM Lcvette: someone is working on it still
06:02 PM Lcvette: ah
06:02 PM Tom_L: not as actively
06:02 PM JT[m]: right
06:03 PM JT[m]: since Jeff left it has slowed way down and got very bloated with GUI's
06:03 PM Tom_L: but doing it with gcodes would be a bad idea imo even if it were possible
06:03 PM JT[m]: doing what?
06:03 PM Tom_L: writing to the file
06:03 PM Tom_L: it was mentioned....
06:04 PM Lcvette: i saw that if there is a power outage you loose the var file if it hasn't saved
06:04 PM JT[m]: ohh you don't want to write to the var file
06:04 PM Tom_L: do you know when that gets written back to the drive?
06:04 PM JT[m]: here's a thought the time stamp could be used to know when it was last written to...
06:05 PM JT[m]: no clue, just know it's not instant ALL the time
06:05 PM Tom_L: i figured that
06:06 PM Tom_L: Lcvette so you set the offsets again :)
06:06 PM JT[m]: there is always a backup var file .var.bak
06:08 PM Tom_L: huh, i didn't know that
06:08 PM Tom_L: i see the file
06:09 PM JT[m]: it seems to be a copy of the .var file at startup and is not updated at shutdown
06:10 PM Tom_L: maybe if it's not there it gets created on startup
06:10 PM Tom_L: and once it's there it stays
06:11 PM Tom_L: if i undestand the code right, i think that example reads the current offset values from memory
06:11 PM JT[m]: looks like it updates both files when something is changed that is in the .var file
06:11 PM Tom_L: that's how it should be
06:11 PM JT[m]: just watching the time stamp
06:13 PM Lcvette: so it updates?
06:13 PM JT[m]: it?
06:14 PM JT[m]: 2 more days of this heat then nice fall weather
06:14 PM Tom_L: yep
06:14 PM Tom_L: not sure when but it's headed this way
06:14 PM Lcvette: JT wrote "looks like it updates both files when something is changed that is in the .var file
06:14 PM Lcvette: "
06:14 PM JT[m]: friday here and good we are riding the Spyder to Murray Ky for the weekend
06:15 PM Tom_L: might need a jacket
06:15 PM JT[m]: yea I pressed the clear offsets button in axis
06:15 PM JT[m]: yep
06:15 PM JT[m]: for the morning ride Saturday for sure
06:16 PM Lcvette: clear offset that sounds like a good button to have
06:16 PM JT[m]: yea, in axis you can clear any coordinate system offset as well as G92 offsets from the menu
06:17 PM Tom_L: i'd take a look at the axis code and see how they go about it
06:17 PM Lcvette: when turboss returns i will see if we can clear an offset line from the table
06:17 PM * JT[m] thinks hazzy is home from work...
06:17 PM Tom_L: i had it open last night but really didn't know what i was looking at
06:17 PM Lcvette: no idea how to look at the axis code
06:18 PM Lcvette: lol
06:18 PM Tom_L: download the linuxcnc source and open it
06:18 PM JT[m]: don't try, it's a mix of python, tcl, and something else
06:18 PM Tom_L: yeah it's a mess
06:18 PM Lcvette: yeah sounds like a bag of trouble
06:18 PM JT[m]: it's hard to even find all the files lol
06:18 PM Tom_L: what exactly is tcl?
06:18 PM Tom_L: i think that file was tcl
06:19 PM Lcvette: hazzy vanished
06:19 PM JT[m]: a complicated language that only Dewey understands lol
06:19 PM JT[m]: https://www.tcl.tk/about/language.html
06:20 PM Tom_L: is he one of the few active developers left?
06:20 PM JT[m]: looks like a mix of basic with the $ all over the place
06:20 PM JT[m]: yea from time to time he fixes something
06:21 PM Tom_L: i see him pop in irc once in a while briefly
06:21 PM Lcvette: you mentioned linuxcnc is bloatd with guis
06:22 PM Lcvette: can you not just delete the gui files?
06:22 PM JT[m]: I see him on the forum from time to time
06:22 PM JT[m]: too intertwined lol
06:22 PM Lcvette: oh
06:22 PM JT[m]: one of my projects is/was to remove all gui's from linuxcnc
06:22 PM Lcvette: yeah?
06:23 PM Lcvette: got too compicated?
06:24 PM JT[m]: problem I ran into is trying to sort out the startup script... it's in hieroglyphics
06:24 PM Lcvette: :o
06:24 PM JT[m]: not really but a complicated bash script, I sorted out most of it a while back
06:24 PM Lcvette: they went egyption on ya
06:24 PM Tom_L: JT[m], so when you clear the offset does that just clear it in memory then later linuxcnc sees it and writes the changes out to the file?
06:25 PM Lcvette: $%$^$^&%&*(^)^*&%(%&(^%*)&
06:25 PM Lcvette: lol
06:25 PM Tom_L: kindof a housekeeping chore for it...
06:25 PM JT[m]: I think it's usually pretty instant but there are some cases where it's not
06:25 PM Lcvette: o/
06:25 PM JT[m]: here comes hazzy
06:25 PM Lcvette: nubbyb has arrived
06:25 PM hazzy-m: o/
06:26 PM Tom_L: but it _does_ change the copy stored in memory
06:26 PM Lcvette: \o/
06:26 PM hazzy-m: JT was right! just got home and read the log
06:27 PM JT[m]: Tom_L: yes I'm sure the variables in memory are updated instantly
06:27 PM hazzy-m: I think Jeff is responsible for all the TLC, he wrote a TLC based GUI editor at one of the places he worked and was hoping to use it with LCNC eventualy
06:27 PM JT[m]: IIRC the issue is when does it write those values to the .var file
06:27 PM Tom_L: then some time later it gets written out to file
06:27 PM Tom_L: yeah
06:27 PM JT[m]: I think your right about that hazzy
06:28 PM JT[m]: does riot have a speall checkr
06:28 PM Tom_L: so blanket changing the whole offset table and writing it may be a bad idea
06:28 PM Lcvette: not that im aware of
06:29 PM JT[m]: Tom_L: exactly
06:29 PM Tom_L: gotcha
06:29 PM Tom_L: i don't fully undestand/follow the code
06:29 PM Lcvette: hazzy are you close on the persistent boxes?
06:29 PM hazzy-m: the strange ting about the var file is I've not had good luck watching it with file system watcher, if never pickes up the changes, I think maybe the time stampes or other metatada is copied over when it 's writen
06:29 PM Lcvette: i watch it in sublime
06:30 PM Lcvette: thats where i saws the offsets updating
06:30 PM hazzy-m: JT: no, it really neads one LOL
06:30 PM * hazzy-m can hardly spell his own name
06:31 PM Lcvette: "Lcvette
06:31 PM Lcvette: hazzy are you close on the persistent boxes?"
06:39 PM hazzy-m: getting there
06:39 PM JT[m]: got my federal refund and it was not what I filed for...
06:39 PM Lcvette: More or less?
06:39 PM JT[m]: I filed for $2,200 but the check was for $4,100
06:39 PM Tom_L[m]: at least they haven't come lookin fer ya
06:40 PM Lcvette: Hurray!
06:40 PM JT[m]: must have missed a quartly payment
06:40 PM JT[m]: ah it's my money all the time they just borrowed it
06:40 PM JT[m]: Tom_L: yea
06:41 PM JT[m]: just got to live through the month to get my first goverment check!
06:41 PM Tom_L: yay
06:41 PM Tom_L: they'll tax ya on that too...
06:41 PM Lcvette: hazzy: do you have a time frame?
06:41 PM JT[m]: yep you have to pay taxes on it
06:42 PM Tom_L: ss too?
06:42 PM Tom_L: or just fed?
06:42 PM JT[m]: ss
06:42 PM JT[m]: fed I already paid that tax
06:42 PM Tom_L: i mean on the ss income
06:43 PM JT[m]: yea ss income is taxable
06:43 PM Tom_L: sux
06:43 PM JT[m]: well you don't pay taxes on what you pay in but you do on what you draw out
06:43 PM Tom_L: right
06:44 PM JT[m]: so yea it sucks to pay the higher rate 40 years later
06:45 PM JT[m]: opps 50 years later lol
06:51 PM Tom_L: https://www.youtube.com/watch?v=0KRT88gZQR0
06:51 PM Tom_L: today
06:51 PM jthornton: what's that?
06:51 PM Tom_L: taiwan bridge collapse
06:51 PM jthornton: rut row
06:52 PM Tom_L: tanker truck didn't quite make it across
06:53 PM jthornton: yikes, he had a bad day
06:57 PM * jthornton watches chicken tv
07:33 PM hazzy-m: ouch, need to get AVE doing the falure analysis on that
07:53 PM Tom_L: kinda looks like a cable popped
07:53 PM Tom_L: hard to tell
11:02 PM Lcvette: \o/