[Hamlib-developer] Broken pipe causes crash after closing flrig

[Philip R. <gm...@bt...>]

I have a situation where my app crashes if I close flrig before I close 
my app. I was deliberately doing this as i wanted to switch to native 
hamlib access of the rig (IC-7300) for testing a new feature in my app.

In my app I poll hamlib and thus flrig for frequency, mode etc every 
second. When I do this after closing flrig, I get the following stack:

Thread 34 "zzalog" received signal SIGPIPE, Broken pipe.
[Switching to Thread 0x7ffff0b1d6c0 (LWP 1061563)]
__GI___libc_write (nbytes=236, buf=0x7ffff0b185d0, fd=12)
     at ../sysdeps/unix/sysv/linux/write.c:26
26    ../sysdeps/unix/sysv/linux/write.c: No such file or directory.
(gdb) bt
#0  __GI___libc_write (nbytes=236, buf=0x7ffff0b185d0, fd=12)
     at ../sysdeps/unix/sysv/linux/write.c:26
#1  __GI___libc_write (fd=12, buf=0x7ffff0b185d0, nbytes=236)
     at ../sysdeps/unix/sysv/linux/write.c:24
#2  0x00007ffff62caba5 in write_block
     (p=p@entry=0x7fffdc008358, txbuffer=txbuffer@entry=0x7ffff0b185d0 
"POST /RPC2 HTTP/1.1\r\nUser-Agent: XMLRPC++ 0.8\r\nHost: 
127.0.0.1:12345\r\nContent-type: text/xml\r\nContent-length: 
119\r\n\r\n<?xml 
version=\"1.0\"?>\r\n<?clientid=\"hamlib(40601)\"?>\r\n<methodCall><methodName>rig.ge"..., 
count=236)
     at iofunc.c:1112
#3  0x00007ffff631aef1 in write_transaction
     (xml_len=<optimized out>, xml=0x7ffff0b185d0 "POST /RPC2 
HTTP/1.1\r\nUser-Agent: XMLRPC++ 0.8\r\nHost: 
127.0.0.1:12345\r\nContent-type: text/xml\r\nContent-length: 
119\r\n\r\n<?xml 
version=\"1.0\"?>\r\n<?clientid=\"hamlib(40601)\"?>\r\n<methodCall><methodName>rig.ge"..., 
rig=0x7fffdc000b70) at flrig.c:545
#4  flrig_transaction
     (rig=rig@entry=0x7fffdc000b70, cmd=cmd@entry=0x7ffff649a6e5 
"rig.get_vfoA", cmd_arg=cmd_arg@entry=0x0, 
value=value@entry=0x7ffff0b1c620 "", value_len=value_len@entry=128) at 
flrig.c:583
#5  0x00007ffff631e047 in flrig_get_freq
     (rig=0x7fffdc000b70, vfo=<optimized out>, freq=0x7ffff0b1c7f8)
     at flrig.c:1221
#6  0x00007ffff62adb6b in rig_get_freq
--Type <RET> for more, q to quit, c to continue without paging--
     (rig=0x7fffdc000b70, vfo=1, freq=0x7ffff0b1c7f8) at rig.c:2571
#7  0x0000555555767c06 in rig_if::th_read_values() (this=0x55555a3ea1d0)
     at src/rig_if.cpp:348
#8  0x0000555555767b55 in rig_if::th_run_rig(rig_if*) (that=0x55555a3ea1d0)
     at src/rig_if.cpp:333


Is it possible to catch the broken pipe and return an appropriate error 
code in "rig_get_freq" (and similar calls)?


Phil GM3ZZA