Hi Fred, the problem should be fixed in https://github.com/dl1jbe/tlf/tree/fix_bandmap
I runs now stable with N8NOEs 7373 port with around 600 active spots. 73, de Tom DL1JBE Am Tue, 2 Feb 2021 22:19:01 +0100 schrieb Fred Siegmund <dh...@freenet.de>: > Hi Tom, I will test at the weekend in EUDX contest with 'gdb'. > > 73 Fred > > Am 02.02.21 um 21:18 schrieb Thomas Beierlein: > > Hi Fred, > > > > Just saw Zolis test results for the RBN. Normally bandmap seems to > > be very stable even during large contests (at least with normal DX > > cluster). > > > > There is one possible variation to test during next major contest. > > > > Prepare and build tlf as follows and make sure to have the 'gdb' > > debugger installed. > > > > CFLAGS="-g" ./configure --prefix=<your normal prefix> > > > > That will build tlf with debug information built in. Install the > > program like normal and than start tlf with > > > > gdb path/to/tlf <all the switches you normally use> > > > > Tlf should work like before but in case of a segmentation fault > > stop in gdb's command line. > > > > Just do a 'bt' command for getting backtrack information and send us > > the output. you can quit gdb afterwards with 'q'. > > > > 73, de Tom > > > > > > Am Sun, 31 Jan 2021 13:40:10 +0100 > > schrieb Fred Siegmund <dh...@freenet.de>: > > > >> Also something I see from time to time (repeatedly): > >> > >> 73 Fred > >> > >> Am 31.01.21 um 12:58 schrieb Fred Siegmund: > >>> Hi Tom, i install always with --prefix option and start from /bin. > >>> But its acually easier now to edit logcfg.dat, only that it is at > >>> 2 locations was confusing (for me). > >>> > >>> For the bandmap crashes its kind of hard to reproduce, I always > >>> use ARcluster like N8NOE, because of the great filtering options > >>> (much better than with DXspider). I just notice that with high > >>> load (like in major contests) it happens quite often. Right now > >>> just dawdling in REF contest its no problem, I have no crashes. > >>> Maybe there is some way to instrument the code, I just see > >>> segmentation fault if it happens. > >>> > >>> 73 Fred > >>> > >>> Am 30.01.21 um 17:01 schrieb Thomas Beierlein: > >>>> Hi Fred, > >>>> > >>>> Am Sat, 30 Jan 2021 14:19:59 +0100 > >>>> schrieb Fred Siegmund <dh...@freenet.de>: > >>>> > >>>>> Also I noticed that logcfg.dat gets copied to the /bin directory > >>>>> and used from there. Seems to be new. > >>>>> > >>>> maybe by accident TLf got started there? > >>>> > >>>> It normally assumes to be started from a contest working > >>>> directory. If there is no logcfg.dat in that directory old > >>>> versions used the default (readonly) version > >>>> (/usr/(local)/share/tlf/logcfg.dat). Newer version copies the > >>>> default file into the actual directory so you can adapt it to > >>>> your needs. > >>>> > >>>> 73, Tom > >>>> > >>>> P.S. How do you connect to the reverse beacon server? Maybe we > >>>> can reproduce your problems > >>>> > >>>>> 73 Fred > >>>>> > >>>>> Am 27.01.21 um 20:04 schrieb Zoltan Csahok: > >>>>>> @dh5fs <https://github.com/dh5fs> I suggest closing this issue > >>>>>> (if the solution is OK) and opening a new one on further > >>>>>> improvement options. > >>>>>> > >>>>>> — > >>>>>> You are receiving this because you were mentioned. > >>>> > > > > > -- "Do what is needful!" Ursula LeGuin: Earthsea --
