I do not experience this issue when switching profiles. Profile switches are also much quicker than the 3 seconds reported. Suggest that Mike’s case may include a user network issue.
Flex 6600, Mac OS Catalina. WSJT-X 2.1.2 George J Molnar Arlington, Virginia, USA (702) 767-5420 KF2T - FM18lu > On Apr 23, 2020, at 2:08 PM, Bill Somerville <g4...@classdesign.com> wrote: > > Hi Mike, > > so what's wrong with the WSJT-X user pressing Retry when changing a SmartSDR > profile breaks the CAT connection? > > 73 > Bill > G4WJS. > > On 23/04/2020 18:20, Black Michael via wsjt-devel wrote: >> Yes.... >> >> >> >> >> On Thursday, April 23, 2020, 12:19:11 PM CDT, Bill Somerville >> <g4...@classdesign.com> <mailto:g4...@classdesign.com> wrote: >> >> >> Hi Mike, >> >> Does this cause WSJT-X to offer a Retry/Reconfigure/Cancel message box? >> >> 73 >> Bill >> G4WJS. >> >> On 23/04/2020 18:00, Black Michael via wsjt-devel wrote: >>> The Flex takes almost 3 seconds to do a profile change and with a polling >>> rate of 1 second the code in PollingTransceiver.cpp times out and does a >>> shutdown so sends "q" to rigctld for example. >>> >>> Not talking about changing the polling_rate....need to add a counter to >>> only time out after 4 seconds instead of using the >>> polling rate to determine the timeout. >>> >>> Any rig should be able to poll at 1 second (or even faster for that matter) >>> but the timeout should not be 1 second....it needs to be the worst case >>> scenario to avoid WSJT-X from disconnecting unnecessarily. >>> >>> Mike >>> >>> >>> >>> >>> >>> >>> >>> >>> On Thursday, April 23, 2020, 11:49:27 AM CDT, Bill Somerville >>> <g4...@classdesign.com> <mailto:g4...@classdesign.com> wrote: >>> >>> >>> On 23/04/2020 17:34, Black Michael via wsjt-devel wrote: >>>> Found a problem with Flex which is fixed by changing the rig polling rate >>>> to 3 seconds and adding some changes to hamlib. >>>> Doing a profile change on Flex can take almost 3 seconds. >>>> So if you have polling at 1 second WSJT-X times out waiting for the >>>> profile change and actually requests hamlib to disconnect *(i.e. shutdown). >>>> This may explain random disconnects when any rig takes too long to respond >>>> on any command. >>>> >>>> There are a few possible solutions which I'll rank in my preference >>>> order.... >>>> >>>> #1 Fix the polling to provide a fixed timeout of 4 seconds. Users won't >>>> have any idea what to set for this and I don't see any problem with >>>> waiting up to 4 seconds when things are getting delayed. Normally we >>>> don't run into this timeout but we need to gracefully recover when we do. >>>> #2 Provide a user-controllable setting which defaults to 4 seconds. Just >>>> in case this longer timeout causes a different problem. Maybe put it in >>>> the release candidate and then remove it when nobody complains. >>>> >>>> So...for Flex users....change polling to 3 seconds and use this >>>> hamlib_settings.json file in the WSJT-X configuration directory >>>> This changes will be in the next version of WSJT-X for the Flex6xxx entry. >>>> >>>> { >>>> "config": { >>>> "retry": "13", >>>> "post_write_delay": "0", >>>> "timeout": "300" >>>> } >>>> } >>>> >>>> >>>> de Mike W9MDB >>>> >>> Hi Mike, >>> please explain what you mean by requests Hamlib to disconnect? Are you >>> referring to a TCP/IP connection failure? Does this cause WSJT-X to offer a >>> Retry/Reconfigure/Cancel message box? >>> >>> A 3 s polling interval being enforced is not acceptable for an issue that >>> only happens when a profile is changed on SmartSDR. >>> >>> 73 >>> Bill >>> G4WJS. >>> > > _______________________________________________ > wsjt-devel mailing list > wsjt-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/wsjt-devel
_______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
