#linuxcnc-devel Logs

Jun 17 2017

08:04 AM jepler: wheezy is now oldoldstable
12:43 PM seb_kuzminsky: our project is like a dinosaur and rtai is the tar pit we're trapped in
04:04 PM KGB-linuxcnc: 03Sebastian Kuzminsky 052.7-linux4.4 5768124 06linuxcnc 10src/hal/drivers/pluto_step.comp pluto: use rtapi's fabs() instead of the kernel's abs() * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5768124
04:21 PM mozmck: I am looking at using SPI with the hostmot2 bspi driver. I'm curious why there is a separate read and write function when SPI is full duplex?
04:32 PM linuxcnc-build_: build #4974 of 1202.rip-lucid-amd64 is complete: Failure [4failed compile runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1202.rip-lucid-amd64/builds/4974 blamelist: Sebastian Kuzminsky <seb@highlab.com>
04:35 PM JT-Shop: I'm trying to build dewey's external offset branch on Linux Mint 17.3 with uspace and I get a strange configure error checking for glib... configure: error: no -- required until somebody makes glib optional
04:36 PM JT-Shop: I've tried to apt-get install glib but it is not found
04:36 PM JT-Shop: anyone have a clue?
04:36 PM seb_kuzminsky: JT-Shop: on debian stretch i have libglib2.0-0
04:37 PM seb_kuzminsky: try "apt-cache search libglib"
04:37 PM JT-Shop: thanks
04:40 PM JT-Shop: sigh it says that libglib2.0-0 is the latest and nothing was installed...
04:57 PM linuxcnc-build_: build #4991 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4991 blamelist: Sebastian Kuzminsky <seb@highlab.com>
05:02 PM seb_kuzminsky: that build failure is another Task timeout
05:02 PM seb_kuzminsky: task: 260 cycles, min=0.000016, max=21.993429, avg=0.086069, 2 latency excursions (> 10x expected cycle time of 0.001000s)
05:07 PM JT-Shop: I guess I'll give it a go buiding a new drive with RTAI and see if I can finger out how to get everything working
05:16 PM pcw_home: mozmck: probably because the normal processing model for LinuxCNC is "read --> process --> write" so the HM2 functions support this
05:16 PM pcw_home: and the obligate full duplex nature of SPI is often just in the way
05:20 PM JT-Shop: seb_kuzminsky: is stretch the livecd?
05:22 PM jepler: mozmck: the hm2 driver needs to collect the data to be written, then do the read/write operation, then return the data that was read
05:27 PM pcw_home: Also some slaves are read only or write only ( common D-A, A-D chips for example )
07:33 PM seb_kuzminsky: JT-Shop: no, the current livecd is wheezy
07:55 PM KGB-linuxcnc: 03Sebastian Kuzminsky 052.7 ffd3ccf 06linuxcnc 10src/hal/user_comps/hy_gt_vfd.c hy-gt-vfd: remove some unused variables * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ffd3ccf
07:59 PM jepler: seb_kuzminsky: the fabs/scale compile error is a real bug too
08:18 PM seb_kuzminsky: i thought so, but i wasn't sure
08:35 PM mozmck: pcw_home, jepler, thanks. so can I do both in the read function? It looks like I should be able to.
08:58 PM pcw_home: You would probably have to look at the code (but you have to write _something_ to do a SPI read)
09:47 PM KGB-linuxcnc: 03Jeff Epler 052.7 996d682 06linuxcnc 10docs/man/man3/hm2_bspi_set_read_function.3hm2 10docs/man/man3/hm2_bspi_set_write_function.3hm2 docs: improve some hm2_bspi manpages * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=996d682