Hi All:

Hello from Vancouver Island, Canada!

I originally sent this message yesterday, but I do not know if it made it to 
the list as I got no replies or acknowledgements.  Therefore, here is a repeat 
just to make sure that the message “made it!”

I am running wsjtx-2.7.0-rc6 on an Intel Mac.  My MacOS is version 13.5.2.  I 
have noted several crashes while in MSK144 mode on 6m, over approximately the 
past week.  WSJT-X just disappeared!  To glean more information on the reason 
for the crash I now start the application via the command line.  I do get to 
see error messages in that way.

A typical error message is:

At line 3 of file /Users/jmn/G4KLA/RevXGA2.7.0-rc6/wsjtx/lib/hspec.f90
Fortran runtime error: Actual string length is shorter than the declared one 
for dummy argument 'line1' (-3381519939318841264/80)

I also see similar messages indicating that the issue was in line 2 of 
hspec.f90, this time with an expected string length of 12 and an insane 
negative value for the passed length of the string.

This problem looks very similar to the bunch of errors that were fixed about a 
month of go also related to invalid passed lengths of string parameters.  I 
recall that the earlier errors were caused by miscommunication between a C++ 
caller and a FORTRAN subroutine, and had to do with hardcoded buffer lengths 
between the caller and the invoked routine.  The fix was to call the .f90 
routine with strlen() or sizeof() the buffer parameter and to use this value in 
the .f90 code.

Has anyone else observed this behaviour?  It is repeatable on my machine so I 
can provide more info as necessary.  My Mac is 2019 and admittedly struggles a 
bit with Ventura due to OS bloat, but generally WSJT-X seems to run fine in 
non-MSK144 modes.

73,

Kevin VE7ZD/KN7Q


_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel

Reply via email to