use signalfd to stop on SIGTERM

Is this an overkill? I don't know.

The thing is, correctly intercepting SIGTERM (also SIGINT, etc.) is
incredibly tricky. For example, before this commit, my I/O loops in
server.c and worker.c were inherently racy.

This was immediately obvious if you tried to run the tests. The tests
(especially the Valgrind flavour) would run a worker, wait until it
prints a "Waiting for a new command" line, and try to kill it using
SIGTERM. The problem is, the global_stop_flag check could have already
been executed by the worker, and it would hang forever in recv().

The solution seems to be to use signalfd and select()/poll(). I've never
used either before, but it seems to work well enough - at least the very
same tests pass and don't hang now.

1 files changed, 5 insertions, 1 deletions

diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index af6e9fb..fb94b3a 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -39,6 +39,7 @@ generate_sql_header(sqlite)
 add_my_executable(server server_main.c server.c
     cmd_line.c
     command.c
+    event_loop.c
     file.c
     msg.c
     net.c
@@ -48,6 +49,7 @@ add_my_executable(server server_main.c server.c
     sqlite.c
     storage.c
     storage_sqlite.c
+    string.c
     tcp_server.c
     worker_queue.c)
 target_link_libraries(server PRIVATE pthread sqlite3)
@@ -63,11 +65,13 @@ add_my_executable(worker worker_main.c worker.c
     ci.c
     cmd_line.c
     command.c
+    event_loop.c
     file.c
     git.c
     msg.c
     net.c
     process.c
     run_queue.c
-    signal.c)
+    signal.c
+    string.c)
 target_link_libraries(worker PRIVATE git2 pthread)