#linuxcnc-devel Logs

Oct 26 2018

#linuxcnc-devel Calendar

10:11 AM mozmck: I'm pleased with the results of having the first post of new forum users moderated. I don't believe any spam has made it onto the forum, but I've killed a good bit in moderation. It is easier to delete the spam posts because they are all on the moderation page, and the workload is overall lower.
11:54 AM JT-Shop: cool
02:12 PM andypugh: Yes, I concur. I have tried a test .comp compiled in my root folder with an adjacent include. It compiles without errors in 2.7 and master. However I know I have had problems in the past, and forgot to check that compile failed without the patch, simply checking that it passed with the patch.
02:12 PM andypugh: I really should know better.
02:20 PM andypugh: jepler_: https://forum.linuxcnc.org/24-hal-components/35418-halcompile-missing-additionnal-include-file#119326 is all the context I have
02:21 PM andypugh: I can confirm that wj200_vfd.comp does fail to compile with halcompile, but that is it not finding <modbus.h>
02:34 PM jepler_: that's an example of a userspace component, I guess, and everything I've diddled with has been RT
02:35 PM jepler_: libmodbus-dev: /usr/include/modbus/modbus.h
02:35 PM jepler_: src/config.status:S["LIBMODBUS_CFLAGS"]="-I/usr/include/modbus"
02:35 PM jepler_: so for internal modules we detect this and it ends up helping that .comp to be built in-tree
02:36 PM jepler_: I don't think anything proposed in #515 fixes wj200_vfd.comp to build out of tree
02:37 PM jepler_: I don't care to do more guessing about this, though I feel like #516 *IS* improving something that's worth doing. Assuming "we" believe that these are public, stable APIs of hostmot2.
02:38 PM jepler_: I'm not trying to make anyone commit to that; someone who considers themselves a mostmot2 maintainer would need to have an opinion
02:39 PM jepler_ is now known as jepler
03:29 PM andypugh: jepler: I created the uart and bspi APIs and I do consider them public and stable.
03:29 PM andypugh: (And seem to be the main Hostmot2 maintainer at the moment)
03:43 PM rene_dev_: pcw_home I need firmware with working sserial for the ecm1 board... do you have that?
04:03 PM pcw_mesa_: rene_dev: not sure but I should be able to cobble one together
04:04 PM pcw_mesa_: you also need a V1.5 board or need to patch the RS-422 stuff on a earlier version
04:04 PM rene_dev_: It’s the 1.5 version
04:10 PM rene_dev_: Thanks
04:16 PM pcw_mesa_: are you sure Tormachs configs dont work?
04:21 PM pcw_mesa_: Theres one funny thing about the ECM boards vs Mesa RS-422 boards (inverted RS-422 TXEN)
04:24 PM pcw_mesa_: I can patch the source and invert this, probably the best solution is to have both TXEN polarities available from sserial and connect the desired one in the .ucf file
04:26 PM pcw_mesa_: You might check if the TXEN polarity is wrong on Tormachs config and if so I will do the right thing and patch sserial and hostmot2
04:28 PM pcw_mesa_: (all of our stuff assumes TXEN out of the FPGA is active low since we have all high outputs at startup and dont want TXEN enabled then)
04:31 PM rene_dev_: Tormachs config doesnt work
04:32 PM rene_dev_: does it even use txen? Isn’t that always 1?
04:52 PM pcw_mesa_: The schematic has a pull down so it would probably default to disabled
04:54 PM pcw_mesa_: looking at the .ucf file its just a GPIO bit so woud need to be set to output mode and driven high to enable TX
05:25 PM pcw_mesa_: After I get some things tested here I'll work on a solution to the txen polarity issue (getting pushed around by the boss)
05:30 PM mozmck: ERROR: index out of range, 4 not in [0..4] (increase num_dio/EMCMOT_MAX_DIO=64)
05:34 PM rene_dev_: pcw_mesa no hurry...
07:14 PM pcw_mesa_: rene_dev_: http://freeby.mesanet.com/ecm1_2encs.bit
07:15 PM pcw_mesa_: I fixed it in a decent way so it can be included in the standard source
07:18 PM pcw_mesa_: (different pin tag in the ucf file for active high TXEnables)
09:08 PM commavir_ is now known as commavir
09:37 PM rene_dev_: pcw_mesa thanks, Its working with their bitfile as well now. They seemed to have figured it out.
09:43 PM rene_dev_: Now I need to backport floats to 2.7...