#linuxcnc-devel Logs

Jul 09 2022

01:11 AM pere: andypugh: I suspect the way I set it up I could jog between T1 and M6
02:53 AM pere: I've tried to improve the po4a html parsing, as it is currently horrible. Do not understand why po4a do not like our html pages, as it is perfectly able to handle its own test cases. <URL: https://github.com/mquinson/po4a/pull/366 > added to verify tables work.
02:59 AM -!- #linuxcnc-devel mode set to +v by ChanServ
03:31 AM pere: I finally found the problem with html parsing. we tried to parse it as a manual page...
03:55 AM pere: as hans is offline these days, I hope someone else can commit the translation fixes. The latest one is <URL: https://github.com/LinuxCNC/linuxcnc/pull/1812 >.
05:33 AM andypugh: pere: Yes, you probably could jog between Tn and M6. You probably shouldn’t :-)
07:17 AM pere: I would like to disable the changer when the back door is opened, and disable the jog button when the machine is running on automatic. believe I solved the latter, but noty yet the former.
07:18 AM pere: andypugh: thanks for accepting #1812. Any chance you can generate updated POT and PO files and commit this too?
07:20 AM andypugh: pere:you can disable the jog butons with HAL logic. I would probably use a lut5 component for that, just because it makes it easier to reconfigure the overrides if you want to add more conditions later.
07:20 AM pere: andypugh: yes, used a not + and setting to disable it.
07:21 AM * pere checks out lut5
07:21 AM andypugh: Probably one LUT5 per jog button, it’s a pretty cheap function so one-per-jog saves AND2-ing the physical buttons with the LUT5 output
07:24 AM andypugh: https://docs.google.com/spreadsheets/d/1_VlL39jUOPpS8Tt9mpRK-eaSTS6C6PHBDoE3JitvLzs/edit#gid=0
07:25 AM andypugh: It’s very simple initially, the config in the sheet above gives jog if jog-button & not auto & not back panel
07:25 AM andypugh: You could also add spindle-not-turning (for example)
07:25 AM pere: how can I make sure M6 stops/hang/waits until the carousel have the correct tool pocket ready?
07:26 AM pere: the machine encorse spindle-not-turning already.
07:26 AM pere: s/encorse/enforce/
07:26 AM andypugh: If you have the spindle signal already then it is trivial to add that to LUT5 for jog, even if it is redundant.
07:28 AM andypugh: In fact, you don’t even need to change the “function” above for that, because it is set up as jog-button & not auto & not back panel & not spare3 and not spare4
07:29 AM andypugh: I think that iocontrol should already block M6 until tool-prepared == true.
07:29 AM pere: ah, ok.
07:43 AM pere: andypugh: do you know how to update the POT and PO files?
07:45 AM andypugh: I don’t. That was going to be my next question.
07:45 AM andypugh: But I assume there is a script to run
07:51 AM pere: make -C src translateddocs; and the commit the update PO and POT file.
07:51 AM pere: andypugh: before you do, I recommend to fetch the latest translations, to avoid conflicts.
07:53 AM andypugh: How do I fetch the latest translations? Just via git or do I need to do somethign at Weblate?
07:53 AM pere: ie accept #1778 and #1791 before updating the POT PO files.
07:53 AM andypugh: OK.
07:53 AM pere: you could do something on weblate, but I just pressedn 'push' so it is done.
07:54 AM pere: I'm locking weblate while you work, to ensure no conflits arrise.
07:55 AM pere: https://hosted.weblate.org/git/linuxcnc/linuxcnc/ is the weblate git repo, if you want to add it as a remote to track it in the future.
08:03 AM pere: andypugh: just ping me when the lock should be released, or if you got access to weblate yourself, feel free to unlock the repo directly.
08:03 AM andypugh: I am waiting for github to finsih its pre-merge tests
08:11 AM pere: andypugh: not really needed, the 'push' I did and triggered the rebuild, did not really change any texts.
08:11 AM andypugh: OK
08:20 AM pere: it is not really obvious from the man page of lut5 how to specify its behaviour.
08:21 AM pere: ie how to set lut5.N.function and how it is different from the lut5.N function.
09:02 AM andypugh: pere: Did you see the spreadsheet?
09:02 AM Tom_L: maybe a note saying the weights are additive
09:03 AM andypugh: Just set it up with the rquired output for each input, and then use the function from the bottom.
09:03 AM Tom_L: it wasn't clear to me the first time i used it
09:03 AM andypugh: If you are a programmer it’s obvious :)
09:03 AM Tom_L: well of course!
09:03 AM pere: andypugh: yes, and found the python 2 script.
09:04 AM pere: andypugh: my point is that the manual page could use some information, not that the information is not found elsewhere.
09:04 AM andypugh: The inputs are interpreted as 5-bit number that is used as an index into the “function” and the output is the value of the bit found there.
09:05 AM Tom_L: no biggie but the table in 2.8 is centered but not in 2.9
09:10 AM Tom_L: the 4th example function indicates they are additive but maybe easily overlooked
09:13 AM Tom_L: as does the 2nd example
09:14 AM Tom_L: you just have to read it carefully
09:24 AM pere: the python scripts converted ok with 2to3. perhaps the wiki should be updated?
09:39 AM pcw--home: What would be needed to back-port the 2.9 hostmot2 driver to LinuxCNC 2.8?
09:55 AM pere: andypugh: any news on the POT/PO update?
09:55 AM JT-Cave: is it possible to get a diff between branches?
09:55 AM andypugh: pere: I don’t know how to make the POT files.
09:56 AM pere: andypugh: I told you an hour ago? make -C src translateddocs
09:56 AM andypugh: I have tried various versions of your provided incantation from various places in the tree
09:57 AM pere: it should run from the top in a configured source tree. if it is not configured, I guess 'cd src; ./configure' should take care of that. what go wrong?
09:57 AM andypugh: andypugh@rm-one:~$ cd linuxcnc-dev/
09:57 AM andypugh: andypugh@rm-one:~/linuxcnc-dev$ cd src
09:57 AM andypugh: andypugh@rm-one:~/linuxcnc-dev/src$ make -C src translateddocs
09:57 AM andypugh: make: *** src: No such file or directory. Stop.
09:57 AM andypugh: andypugh@rm-one:~/linuxcnc-dev/src$ cd ..
09:57 AM andypugh: andypugh@rm-one:~/linuxcnc-dev$ make -C src translateddocs
09:57 AM andypugh: make: Entering directory '/home/andypugh/linuxcnc-dev/src'
09:57 AM andypugh: emc/Submakefile:4: warning: overriding recipe for target '../include/linuxcnc.h'
09:57 AM andypugh: emc/Submakefile:4: warning: ignoring old recipe for target '../include/linuxcnc.h'
09:57 AM andypugh: Reading 206/206 dependency files
09:58 AM andypugh: Done reading dependencies
09:58 AM pere: drop cd src or -C src?
09:58 AM andypugh: Reading 230/230 realtime dependency files
09:58 AM andypugh: Done reading realtime dependencies
09:58 AM andypugh: make: *** No rule to make target 'translateddocs'. Stop.
09:58 AM andypugh: make: Leaving directory '/home/andypugh/linuxcnc-dev/src'
09:58 AM andypugh: andypugh@rm-one:~/linuxcnc-dev$ cd ../docs
09:58 AM andypugh: -bash: cd: ../docs: No such file or directory
09:58 AM andypugh: andypugh@rm-one:~/linuxcnc-dev$ make -C src translateddocs
09:58 AM andypugh: make: Entering directory '/home/andypugh/linuxcnc-dev/src'
09:58 AM andypugh: emc/Submakefile:4: warning: overriding recipe for target '../include/linuxcnc.h'
09:58 AM andypugh: emc/Submakefile:4: warning: ignoring old recipe for target '../include/linuxcnc.h'
09:58 AM andypugh: Reading 206/206 dependency files
09:58 AM andypugh: Done reading dependencies
09:58 AM andypugh: Reading 230/230 realtime dependency files
09:58 AM andypugh: Done reading realtime dependencies
09:58 AM andypugh: make: *** No rule to make target 'translateddocs'. Stop.
09:58 AM andypugh: make: Leaving directory '/home/andypugh/linuxcnc-dev/src'
09:59 AM pere: How strange. the translateddocs target is listed in linuxcnc/docs/src/Submakefile here. did you disable documentation building in the configure step?
10:02 AM andypugh: Maybe that was it, though I thought they were on.
10:02 AM andypugh: Done reading dependencies
10:03 AM andypugh: make: Nothing to be done for 'translateddocs'.
10:03 AM andypugh: make: Leaving directory '/home/andypugh/linuxcnc-dev/src'
10:08 AM andypugh: The documetation.pot file on the HD seems to be identical to the one on github.
10:15 AM pere: must be something messed up with file timestamps. just remove it and generate it again.
10:16 AM pere: instead of translateddocs, you could also use make -C src pofiles, which should do slightly less and run a bit quicker.
10:21 AM JT-Cave: pcw--home, IIRC I "think" you can do a cherry-pick
10:22 AM JT-Cave: https://stackoverflow.com/questions/9339429/what-does-cherry-picking-a-commit-with-git-mean
10:24 AM pere: andypugh: lut5 worked fine. <URL: https://github.com/SebKuzminsky/MazakVQC1540/commit/329c63236d405127075968cf71dbc82be2af8b57 >
10:37 AM pere: silopolis[m]: around? Can you join us in po4a on irc.debian.org to discuss a bug of yours?
10:38 AM andypugh: pere: pofiles did more that translateddocs (the latter said “nothing to do”.
10:39 AM andypugh: I have pushed the new versions, you can unlock weblate.
10:41 AM seb_kuzminsky: pere andypugh : I'm a bit late to the conversation here, but i think it's totally ok to job between Tn and M6
10:41 AM andypugh: seb_kuzminsky: Well, you are likely to not load the programmed tool if you do.
10:42 AM seb_kuzminsky: Tn just tells the toolchanger that the next tool we're going to switch to will be `n`. M6 does the toolchange, including putting the current tool back in its place, moving the toolchanger to the new tool, and picking it up
10:42 AM andypugh: But if you are jogging between Tn and M6 that is likely to be exactly what you want to happen.
10:42 AM pere: andypugh: oh. then I suspect configure disabled docs or translated docs.
10:43 AM andypugh: seb_kuzminsky: Some changers position on the T in preparation for a later M6.
10:44 AM andypugh: So the G-code will typically put the _next_ T on the line after the M6.
10:44 AM seb_kuzminsky: right, that makes good sense if it's the kind of toolchanger that can handle that, like a random toolchange carousel with one of those swapper arms that switch the tool in the spindle with the tool in the currently positioned pocket on the carousel
10:44 AM seb_kuzminsky: but on a machine like the Mazak 15/40, without a swapper arm, that would obviously be wrong, and the machine should not be configured that way
10:44 AM pere: I ended up blocking automatic carousel movement when the door is open, and only allow jog outside auto mode, unless the door is open. hope this will keep fingers safe and the work ok.
10:45 AM pere: andypugh: weblate unlocked.
10:46 AM andypugh: seb_kuzminsky: Then I guess there is a risk of jogging the empty pocket out of position whilst the machine is cutting, and causing a crash when it tries to put the tool away?
10:46 AM pere: yeah
10:46 AM pere: I considered estop if the door is opened during auto mode, but it seemed a little exessive.
10:47 AM seb_kuzminsky: andypugh: you mean jogging the tool magazine while there's a tool in the spindle? yeah, that would be confusing
10:48 AM pere: andypugh: for some strange reason, nb.po do not seem to have the latest files.
10:48 AM seb_kuzminsky: maybe M6 should start by jogging the tool magazine to the pocket of the current tool?
10:48 AM pere: seb_kuzminsky: also, the POT file do not have the latest xhtml fixes either. :(
10:49 AM pere: seb_kuzminsky: probably the safe choice, as it is going to put it away first.
10:50 AM pere: andypugh: you can tell if the POT update was successful by looking at the end. if it show sensible and individual strings, it worked, if it show the E<lt> line noise, it did not.
10:53 AM andypugh: The end of the file or the end of the terminal output?
10:54 AM andypugh: (end of the file looks fine)
10:54 AM andypugh: All I did was blindly merge the two pull requests and attempt to run the commands you said. I really have no idea what I was doing.
10:55 AM pere: 'git diff docs/po/documentation.pot', the end of the file.
10:56 AM pere: when I run 'rm docs/po/documentation.pot ; make -C src pofiles' I get lots of updates to the .pot file. only reason I can think of that you do not is if you are not on master, which seem absurd.
10:58 AM pere: "git log docs/po/documentation.pot" do not show any commit today in my copy of master.
10:59 AM andypugh: No, I don’t think that the “pofiles” option touched anything in docs/po only in src/po
10:59 AM pere: weird. your configure must have disabled it.
11:00 AM pere: andypugh: do you have a new enough po4a installed?
11:00 AM pere: version 0.66 is needed. looking at configure now, there is code to disable translations if po4a is too old.
11:01 AM andypugh: Which configure? There is a new option in debian/configure that I think I saw mentioned, but I don’t think that it relevant here?
11:01 AM pere: src/configure
11:01 AM pere: look for BUILD_DOCS_TRANSLATED
11:01 AM andypugh: I am using —with-realtime=uspace —enable-build-documentation=Y
11:02 AM pere: I see the test is for po4a 0.62. Perhaps the 0.66 issues were worked around in adoc.
11:02 AM pere: what is your po4a --version saying?
11:03 AM Tom_L: some machines have a tool prefetch which would also change the tool pocket
11:04 AM andypugh: pere: I can’t tell you until this build finishes
11:17 AM andypugh: Hmm….
11:17 AM andypugh: po4a: command not found
11:17 AM andypugh: That won’t be heloing
11:41 AM andypugh: Hmm, configure -h lists “--disable-build-documentation-translation” but I don’t think that’s actually the paremeter to use.
11:41 AM andypugh: its enable…. = Y not disable… = N
12:07 PM pere: andypugh: no po4a, no POT generation.
01:29 PM roguish[m]: andypugh: good day. can your lincurv comp be used to vary axes PID parameters based on a variable like axis velocity? I'm having a devil of a time tuning my dc brushed motors and analog drives. I get em tuned at a set speed, and then change velocity and the tune is way off. tune criteria is ferror.
01:30 PM andypugh: roguish[m]: Yes, in fact that is specifically what it was intended for
01:31 PM roguish[m]: oh, cool. I'll give it a try.
01:32 PM roguish[m]: thanks
01:38 PM pere: new po4a release planned in 5 days, to give translators time to work on <URL: https://hosted.weblate.org/projects/po4a/ > and fix the new strings.
02:24 PM silopolis[m]: <pere> "silopolis: around? Can you join..." <- just give me a sec to install an IRC client...
02:28 PM pere: silopolis[m]: thought you already had one, as you are here. :)
02:29 PM silopolis[m]: pere: fallen into the Matrix 😉
02:51 PM -!- #linuxcnc-devel mode set to +v by ChanServ
04:08 PM andypugh: pere: I thought I had it, but it seems not:
04:08 PM andypugh: cd ../docs; po4a -v --msgmerge-opt='-v' --no-update po4a.cfg
04:08 PM andypugh: NOT updating po/documentation.pot as requested (--no-update).
04:08 PM andypugh: src/es/README.adoc doesn't need to be updated.
04:08 PM andypugh: src/es/INSTALL.adoc doesn't need to be updated.
04:23 PM andypugh: pere: By forcing that (i, deleting it in the Makefile) I have new .po and .pot files for docs. Do you want me to merge them?
04:28 PM pere: I would love to have updated POT and PO files in master.
04:29 PM pere: andypugh: look like the documentation.pot dependencies could use some more work. :)
04:30 PM pere: I just remove docs/po/documentation.pot to force POT and PO updates.
04:38 PM andypugh: OK, pushed
04:41 PM pere: andypugh: thank you.
04:55 PM Tom_itx is now known as Tom_L
11:51 PM -!- #linuxcnc-devel mode set to +v by ChanServ