Use sigaction(2) instead of deprecated signal(2), allowing SEGV et al to be
received in handler and therefore in re-spawned server.
Receive SIGSEGV et al in a separate handler function which kills the server
with the same signal if it does not re-spawn.
Log a GDB backtrace on SEGV et al before exiting or respawning.
Rename config option "serval.respawn_on_signal" to "server.respawn_on_crash".
Servald starts DNA helper, receives startup ACK, sends requests, receives
responses, handles malformed helper responses, echoes dnahelper stderr lines to
log, sends MDP reply packet, waits for dead helper process, all asynchronously.
Shuts down helper process during servald shutdown.
Remaining issues:
- Does not impose a timeout on helper responses.
- Only the first URI is reported by the "dna lookup" command.
Was not transmitting actual HTTP server port in rhizome announcements, was
always transmitting port 4110.
When trying for a free HTTP server port, sometimes bind() succeeds but listen()
fails with EADDRINUSE, so new logic to deal with that.
Now write_str(), write_nonblock(), write_all(), set_nonblock() etc. report the
file/line/function of their caller, rather than the function in net.c. Done
using macros, similar in style to WHY() etc.
SET_NONBLOCKING(), SET_BLOCKING(), WRITE_STR() are now set_nonblock(),
set_block() and write_str() respectively, all of which log an error before
returning -1. There are other useful methods: write_all() treats anything less
than all bytes written as an error; write_nonblock() treats EAGAIN and EINTR as
zero bytes written, and a combination: write_all_nonblock().
Fix endless recursion if error or debug logged while reading config file
Fix 'config del' logic
Log messages made before log file can be opened are buffered and written
once the file is open
Do not log to file in ANDROID version, just to Android's log system
... as part of working on work on DNA helper callout facility.
(DNA replies now contain a token that can be used to match them
against DNA requests as a happy side-effect).