Vom: Mon, 2 Feb 2026 22:20:29 +0100 > On 2/2/26 10:15 PM, Daniel Hiepler wrote: > >> [...] caused the existing test to fail [...] > > Sorry got that wrong. Should be "caused the new test to fail". > > (The test included more rsh commands than the existing one. Some > > commands suffered from a race condtion that caused the test to fail > > from time to time.) > Yeah, we are addressing the race conditions. The code is written > multi-threaded, but it lacks locks around basically everything. Pressing > a bit in tests will expose them. Other programs have the same or similar > problems. > > Basically, I've been rewriting linuxcncrsh's code base to get that fixed :-) >
I doubt the cause for the race conditions (at least in my case) are in emcrsh.cc or shcom.cc. IIRC, I found emcCommandWaitDone() [1] basically useless for some commands as those return RCS_STATUS::DONE although they've not updated the model, yet. But I didn't dig further back then. [1] https://github.com/LinuxCNC/linuxcnc/blob/master/src/emc/usr_intf/shcom.cc#L273 Best regards -- Daniel _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
