#linuxcnc-devel Logs

Feb 14 2017

#linuxcnc-devel Calendar

06:07 AM jthornton: is it a bug that C style multi-line comments are not supported by halcompile --document?
06:08 AM archivist: iirc it is a simple parser so no
06:08 AM jthornton: the manual says it works so it must be so :)
07:31 AM KGB-linuxcnc: 03John Thornton 052.7 d82230b 06linuxcnc 10docs/src/hal/comp.txt Docs: add info about multi-line comments and add info for the --install option * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d82230b
07:32 AM jepler: jthornton: uhhhhhhhh
07:32 AM jepler: triple quotes do NOT introduce comments
07:32 AM jepler: they introduce (doc) strings
07:32 AM jthornton: ok
07:33 AM jepler: so for instance if you write this: pin out bit out /* this is really embarassing */ ;
07:33 AM jepler: nothing goes in the docs, because /* ... */ is a comment
07:33 AM jepler: if you write: in out bit out "this is really embarassing"; // or the triple-quoted variety
07:33 AM jepler: then the string "this is really embarassing" will go in the documentation and you will be really embarassed
07:34 AM jthornton: so the example I looked at had /* some text */ in the header
07:34 AM jepler: (A python-style comment is introduced with "#" and continues to the end of the line, but halcompile doesn't use the # character this way)
07:34 AM jepler: what example are you looking at?
07:34 AM jthornton: stmbl.comp
07:35 AM jepler: is that a file in the linuxcnc source tree?
07:35 AM jthornton: http://paste.ubuntu.com/23994807/
07:35 AM jthornton: to be honest I don't know where I got it
07:35 AM jepler: OK what about it
07:36 AM jthornton: can you revert that commit?
07:36 AM jepler: you are asking me to revert that commit?
07:36 AM jthornton: yes
07:37 AM KGB-linuxcnc: 03Jeff Epler 052.7 8266031 06linuxcnc 10docs/src/hal/comp.txt Revert "Docs: add info about multi-line comments and add info for the --install option" * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8266031
07:38 AM jepler: afk, need to get ready for work
07:38 AM jthornton: I sent myself a note about the --install needed sudo and tested the other options on the first comp I could find this morning sorry
07:38 AM jthornton: me to
07:38 AM jthornton: thanks for catching that and fixing it
07:38 AM jepler: halcompile --install needs sudo *FOR INSTALLED VERSIONS*
07:39 AM jthornton: ok
07:40 AM jepler: what matters is whether the directory where components go is a system directory (the case for installed / packaged versions) or a directory owned by the current version (the case for RIP versions, at least as long as there's only one login)
07:40 AM jepler: er current version = current user
07:41 AM jepler: can you go double check that github issue you opened, and close it if it is mistaken like I think it may have been?
07:41 AM jthornton: yes
07:44 AM jthornton: ah yes the example I looked at was backwards ....
08:28 AM * JT-Shop hopes he is done making blunders today
08:29 AM jepler: making good software is just the discipline of making more correct choices than blunders overall
09:00 AM linuxcnc-build: build #748 of 4017.5.deb-wheezy-armhf is complete: Failure [4failed shell_3] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4017.5.deb-wheezy-armhf/builds/748 blamelist: John Thornton <bjt128@gmail.com>
09:52 AM seb_kuzminsky: oh wheezy-armhf, what am i going to do with you?
09:55 AM archivist: make its failure not be the end of the world for other platforms
10:04 AM jepler: if that's feasible to code in the buildbot configuration language, I think it would be a decent solution for now
01:38 PM seb_kuzminsky: i mailed the SoftIron folks a week ago, they said they'd get back to me within 48 hours, 48 hours later i got an automatic email saying they'd get back to me later, then nothing since then
01:39 PM jepler: gughg
01:50 PM jepler: scaleway.com offers bare metal ARM servers for $2.99/month. "The C1 server has a 4-cores ARMv7 CPU with 2GB of RAM and a 1 Gbit/s network card." Processor : Marvell PJ4Bv7 Processor rev 2 (v7l) BogoMIPS : 1332.01
01:52 PM cradek: wow that's about 1.5% as fast as the machine I'm typing on
01:53 PM skunkworks: now now...
01:57 PM cradek: no seriously it sounds smart, if someone else has to deal with the overheating and other hardware stuff
01:59 PM jepler: Let me time building linuxcnc on one...
02:00 PM seb_kuzminsky: is that price in addition to the Business Service at 49 Euros/month? or is Business Service an optional add-on?
02:01 PM jepler: hope I didn't miss some fine print. I signed up as an individual and they preauthed my card for just 2.00EUR
02:03 PM jepler: yes those are additional levels
02:04 PM jepler: 2.99EUR/month gets you "developer" which promises a <12 hour response time on tickets, then there are business and enterprise which offer higher SLA, better response time on tickets, etc
02:04 PM jepler: (oh and it's 2.99EUR/month, I said $ the first time.)
02:14 PM jepler: is it a bad sign that I've already gotten bored waiting for build-deps to install?
02:18 PM seb_kuzminsky: the breakeven point between their monthly and hourly rate is 500 hours/month, which is ~16 hours/day
02:18 PM cradek: so 2 builds :-)
02:18 PM seb_kuzminsky: if you use the machine less than that it's worth it to pay by the hour
02:19 PM seb_kuzminsky: heh
02:22 PM jepler: I wonder, with 2GB RAM, whether I should try building -j2 or -j4 .. there are 4 cores
02:22 PM jepler: oh here's one reason it's slow: the root filesystem is mounted over nbd
02:26 PM jepler: having your data on an SSD doesn't help much if you add 2-4ms to every I/O request for network roundtrips :(
02:29 PM seb_kuzminsky: jepler: on my 4-core, 2 GB, swapless Odroid U3 i build with -j6
02:32 PM jepler: seb_kuzminsky: thanks for that data point .. maybe that's why it melts?
02:34 PM jepler: installing build-depends done at 34 minutes uptime
02:34 PM jepler: now starting a build that *EXCLUDES* documentation but includes the testsuite, just like on travis-ci..
02:35 PM seb_kuzminsky: yeah maybe i should try -j1
02:35 PM seb_kuzminsky: the buildbot builds the docs on every supported platform, for better or worse
02:47 PM jepler: -j4 is modestly into the 2GB swap file I added (20M swapped out)
02:50 PM jepler: running tests at 51 minutes uptime
02:53 PM jepler: (so that's only about 17 minutes build time)
02:54 PM seb_kuzminsky: nice
02:55 PM seb_kuzminsky: the U3 spends 19:45 in make (which includes docs)
02:55 PM seb_kuzminsky: then 11:28 in runtests
02:55 PM KimK_laptop: I accidentally discovered that the forum is down. Sorry if I'm the bearer of bad news, hope it's because someone is working on it.
02:56 PM jepler: ssh login doesn't complete within a reasonable length of time, power cycling..
02:57 PM jepler: thanks KimK_laptop
02:58 PM jepler: forum is loading for me now
03:01 PM jepler: security updates, including php security update, on forum. apache server restarted again.
03:04 PM seb_kuzminsky: thanks jepler
03:07 PM jepler: hm do we have the right modal codes in linuxcnc to do pourover coffee? https://hackaday.io/project/2543-debrew
03:09 PM seb_kuzminsky: linuxcnc-build: force build --branch=master 0000.checkin
03:09 PM linuxcnc-build: The build has been queued, I'll give a shout when it starts
03:09 PM jepler: tests finished at uptime 1:04
03:09 PM jepler: so 13 minutes in runtests
03:10 PM jepler: now timing doc build
03:11 PM seb_kuzminsky: so that's within a few minutes of a U3, that's great
03:11 PM linuxcnc-build: build #4770 forced
03:11 PM linuxcnc-build: I'll give a shout when the build finishes
03:11 PM seb_kuzminsky: the cloud is not where i look for hardware, i must be falling behind the times
03:12 PM jepler: it's also a matter of attitude, whether you are predisposed to like depending on some nebulous online service instead of administering your own hardware
03:13 PM jepler: (though this company initially made its splash by promoting the bare-metal ARM servers, only one of their many configurations is ARM, and it's the lowest-end one)
03:14 PM seb_kuzminsky: it's almost as if arm sucks in the server space
03:14 PM jepler: (so depending on them maintaining it, when they've decided x86_64 is the right way, may not be the greatest idea)
03:14 PM linuxcnc-build: build #2429 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/2429
03:14 PM seb_kuzminsky: derp
03:15 PM seb_kuzminsky: it's all going to fail
03:15 PM linuxcnc-build: build #2913 of 1400.rip-wheezy-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/2913
03:15 PM seb_kuzminsky: prepare for a failvalanche
03:15 PM linuxcnc-build: build #3114 of 1404.rip-wheezy-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1404.rip-wheezy-rtpreempt-amd64/builds/3114
03:15 PM linuxcnc-build: build #4752 of 1300.rip-precise-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1300.rip-precise-i386/builds/4752
03:15 PM linuxcnc-build: build #1381 of 1520.rip-jessie-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1520.rip-jessie-amd64/builds/1381
03:16 PM linuxcnc-build: build #2919 of 1405.rip-wheezy-armhf is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1405.rip-wheezy-armhf/builds/2919
03:16 PM linuxcnc-build: build #1382 of 1530.rip-jessie-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1530.rip-jessie-rtpreempt-amd64/builds/1382
03:17 PM linuxcnc-build: build #4755 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/4755
03:19 PM seb_kuzminsky: ok, now every buildslave runs with nproc parallel make jobs, except wheezy-armhf which runs with 2
03:29 PM jepler: doc building (at -j5) still going..
03:44 PM seb_kuzminsky: with 2 instead of 6 threads making, the U3 is running a lot slower and cooler. no surprises there
03:55 PM jepler: building only docs took 29 minutes
03:56 PM jepler: so right around 1 hour for RIP build + docs + runtests
03:56 PM jepler: data point gathered, disposing of machine..
03:59 PM cradek: hmm.
05:08 PM seb_kuzminsky: a U3 with -j2 builds rip + docs + runtests in ~46 minutes
05:11 PM seb_kuzminsky: linuxcnc-build: force build --branch=master 0000.checkin
05:11 PM linuxcnc-build: build #4771 forced
05:11 PM linuxcnc-build: I'll give a shout when the build finishes
05:21 PM Tom_itx is now known as Tom_L
06:29 PM linuxcnc-build: Hey! build 0000.checkin #4771 is complete: Success [3build successful]
06:29 PM linuxcnc-build: Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4771
10:10 PM KGB-linuxcnc: 03Dewey Garrett 05dgarr/external_offsets fd0ee76 06linuxcnc 10(23 files in 9 dirs) External Axis Offset hal pins * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fd0ee76