#linuxcnc-devel Logs

Mar 17 2018

09:25 AM jepler: pcw_mesa: can you give me a one-sentence summary of what 'SampleLate' does in BSPI?
09:53 AM pcw_home: SampleLate (if true) samples the received SPI data 1/2 SPI clock later than normal, useful when high clock rates or isolation cause significant delays from clock to received data
09:55 AM linuxcnc-build: build #3581 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/3581 blamelist: Jeff Epler <jepler@unpythonic.net>
10:03 AM jepler: Segmentation fault
10:03 AM jepler: make: *** [.html-images-stamp] Error 139
10:03 AM jepler: sigh that odroid hardware
10:07 AM jepler: pcw_mesa: thanks
10:18 AM linuxcnc-build: build #5430 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/5430 blamelist: Jeff Epler <jepler@unpythonic.net>
08:45 PM jepler: /home/jepler/src/linuxcnc/scripts/linuxcnc: line 571: 8092 Segmentation fault $HALCMD stop
08:45 PM jepler: hm well that's not good
08:45 PM jepler: maybe you shouldn't change the HAL ABI and rebuild while test are running, jeff
09:00 PM andypugh: I think there are at least two schools of thought on that.
09:06 PM jepler: hm I need a C language lawyer
09:07 PM andypugh: I need sleep.
09:12 PM jepler: huh, OK, I'm mistaken about how C works ;-P
09:13 PM jepler: I knew it was permitted to compute a pointer one past the end of an array, e.g., char arr[16]; char *ptr = arr + 16;
09:14 PM jepler: .. you often do this in idiomatic C, such as a loop for(char *p = arr; p < arr + 16; p++)
09:14 PM jepler: I believed, but now think I was mistaken, that you could compute a pointer one BEFORE an array, for similar reasons
09:14 PM jepler: e.g., whether you are allowed to do a backwards loop over the same array with for(char *p = arr + 15; p > (arr-1); p++)
09:18 PM jepler: .. the question of whether &arr[16] is equivalent to (arr + 16) for the purposes of the past-end-of-array question is secondary
09:18 PM jepler: gcc treated them differently, clang treated them the same
09:22 PM jepler: (C99 standard 6.5.6 clause 8, if you're following along at home)
09:52 PM jepler: If we took out the current motion planner, I guess it would be a trajectomy
10:08 PM jepler: hmph, gcc's ubsan works in debian stretch, clang's fails like sooo
10:08 PM jepler: ../lib/liblinuxcncini.so.0: undefined reference to `__ubsan_handle_dynamic_type_cache_miss'
10:08 PM jepler: clang: error: linker command failed with exit code 1 (use -v to see invocation)
10:08 PM jepler: hal/user_comps/mb2hal/Submakefile:23: recipe for target '../bin/mb2hal' failed
10:08 PM jepler: discussed at length and over my head here http://clang-developers.42468.n3.nabble.com/fsanitize-undefined-and-shared-libraries-td4029561.html
10:14 PM jepler: and gcc -fsanitize=address fails with /usr/bin/ld: objects/rtapi/uspace_rtapi_app.o: undefined reference to symbol 'dlsym@@GLIBC_2.2.5'
10:14 PM jepler: //lib/x86_64-linux-gnu/libdl.so.2: error adding symbols: DSO missing from command line
10:14 PM jepler: which I haven't found discussed yet
10:16 PM jepler: oh weird how has that not been an error all along? -ldl is missing from the commandline