cimple/src/command.h, branch meson My little CI system (hopefully) http://git.tensin.name/cimple/atom?h=meson 2026-09-21T17:19:05Z remove annoying trailing period from repo URLs 2026-09-21T17:19:05Z Egor Tensin egor@tensin.name 2026-09-21T17:19:05Z urn:sha1:588fd2262c2f65e237a5a4e4858e06b4b87aaae3 finally, perfect .clang-format settings? 2026-06-13T18:16:43Z Egor Tensin egor@tensin.name 2026-06-13T18:16:43Z urn:sha1:d64e7bf3d93649b78231912f37606153e3d30665 switch from tabs to spaces 2026-06-13T17:51:04Z Egor Tensin egor@tensin.name 2026-06-13T17:51:00Z urn:sha1:84262d9f69aa6730c69f464b5b4fb7f6030bf329 It was fun while it lasted, but I do believe that spaces are strictly superior, and clang-format struggles with tabs while using advanced configuration options. unify my .clang-format files 2026-06-12T15:20:08Z Egor Tensin egor@tensin.name 2026-06-12T15:20:08Z urn:sha1:1160cdb21e0b667269299853cb47a4aa88948dc3 switch to egor@tensin.name 2023-12-12T20:29:42Z Egor Tensin egor@tensin.name 2023-12-12T20:29:42Z urn:sha1:82c5f204346def706fdc7ee28a8e2570fb880369 switch to JSON-RPC as message format 2023-07-17T22:38:50Z Egor Tensin Egor.Tensin@gmail.com 2023-07-17T21:03:59Z urn:sha1:e05f02acd583797ceb449fc501d371d45a4293c1 Instead of the weird `struct msg` I had, I switched to the JSON-RPC format. It's basically the same, but has a well-defined semantics in case of errors. event_loop: add event_loop_add_once 2023-06-13T01:31:05Z Egor Tensin Egor.Tensin@gmail.com 2023-06-13T01:31:05Z urn:sha1:1c0481869fc680ea19627d8babec6de9a6784eb7 event_loop: hide the API details 2023-06-12T23:50:24Z Egor Tensin Egor.Tensin@gmail.com 2023-06-12T23:50:24Z urn:sha1:5b58c9fd2880911aa926259d97a396e97f3d4313 use signalfd to stop on SIGTERM 2023-06-12T23:37:08Z Egor Tensin Egor.Tensin@gmail.com 2023-06-11T23:42:08Z urn:sha1:48ce9170b057ddd2165b0239a92aede15849f7a3 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. rework server-worker communication 2023-05-15T13:32:17Z Egor Tensin Egor.Tensin@gmail.com 2023-05-15T13:31:33Z urn:sha1:7cd83e15139447156ca915ce2d9d19295c146d56 OK, this is a major rework. * tcp_server: connection threads are not detached anymore, the caller has to clean them up. This was done so that the server can clean up the threads cleanly. * run_queue: simple refactoring, run_queue_entry is called just run now. * server: worker threads are now killed when a run is assigned to a worker. * worker: the connection to server is no longer persistent. A worker sends "new-worker", waits for a task, closes the connection, and when it's done, sends the "complete" message and waits for a new task. This is supposed to improve resilience, since the worker-server connections don't have to be maintained while the worker is doing a CI run.