aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/src (unfollow)
Commit message (Collapse)Author
2023-06-13server: handle disconnected workers gracefullyEgor Tensin
2023-06-13event_loop: support more event typesEgor Tensin
2023-06-13event_loop: add event_loop_add_onceEgor Tensin
2023-06-13signal: sigprocmask -> pthread_sigmaskEgor Tensin
2023-06-13signal: refactoringEgor Tensin
2023-06-13sanitize #include-sEgor Tensin
2023-06-13signal: remove unused APIsEgor Tensin
2023-06-13event_loop: hide the API detailsEgor Tensin
2023-06-13use signalfd to stop on SIGTERMEgor Tensin
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.
2023-06-12log: refactoringEgor Tensin
2023-06-12log: thread-safe loggingEgor Tensin
2023-06-12log: prefix with thread IDEgor Tensin
2023-06-12command: fix a race conditionEgor Tensin
Modifying cmd_dispatcher fields like that make it inherently unsafe to call cmd_dispatcher_handle_conn concurrently.
2023-06-12sqlite: fix generate-sql-header.pyEgor Tensin
2023-06-11msg: rework some APIsEgor Tensin
2023-06-09worker_queue: fix worker_queue_add_lastEgor Tensin
2023-06-09STAILQ -> SIMPLEQEgor Tensin
2023-06-09msg: add msg_send_from_argv shortcutEgor Tensin
2023-06-09add {file,net}_close as a wrapper to close(2)Egor Tensin
2023-05-29fix realloc usageEgor Tensin
Remember, this is always a mistake: ptr = realloc(ptr, size); You still need to free() the original ptr if realloc fails.
2023-05-17file: rework file_readEgor Tensin
It now increases the buffer size exponentially until it finishes reading the file.
2023-05-15signal: refactoringEgor Tensin
2023-05-15net: drop the _all suffixEgor Tensin
2023-05-15msg: strings, not wordsEgor Tensin
2023-05-15minor refactoringEgor Tensin
2023-05-15signal: refactoring, add comments in tcp_server, etc.Egor Tensin
2023-05-15valgrind: track open file descriptorsEgor Tensin
2023-05-15EINVAL means EINTR also?Egor Tensin
2023-05-15worker_queue: forgot to close file descriptors in worker_destroyEgor Tensin
2023-05-15cmake: disable -Werror for debug buildsEgor Tensin
2023-05-15rework server-worker communicationEgor Tensin
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.
2023-05-14command: adjust order of parameters to handlersEgor Tensin
2023-05-14process: add process_output_dumpEgor Tensin
2023-05-14msg: add functions for one-off communicationEgor Tensin
2023-05-14msg: enforce at least one wordEgor Tensin
Also, move some stuff to net.c where it belongs.
2023-05-13fix unnecessary #include-sEgor Tensin
2023-05-13ci_queue -> run_queueEgor Tensin
Also, some minor refactoring.
2023-05-13cmake: document _GNU_SOURCE usageEgor Tensin
2023-05-13use -std=c17 -WpedanticEgor Tensin
2023-05-13command: refactoringEgor Tensin
2023-05-13best practices & coding style fixesEgor Tensin
* I don't really need to declare all variables at the top of the function anymore. * Default-initialize variables more. * Don't set the output parameter until the object is completely constructed.
2023-05-13add command module to handle request-response communicationsEgor Tensin
2023-05-13ci_queue: fix a broken getterEgor Tensin
Everything was broken starting from the "making struct ci_queue_entry opaque" commit. Damn, I really wish I'd have some kind of automated testing to catch errors like this...
2023-05-12ci_queue: rename a couple of functionsEgor Tensin
2023-05-07cmake: don't install() valgrind.shEgor Tensin
Turns out, I don't really need to install it for the tests.
2023-05-07valgrind: quiet modeEgor Tensin
2023-05-07add tests to run binaries under ValgrindEgor Tensin
This was quite a bit of refactoring in test/; everything should be more maintainable and robust in theory. Also, valgrind.sh was fixed to use exec (so that signals are passed to the underlying process); Valgrind command line options have also been tweaked. ./ci.sh fails now, but that should be fixable.
2023-05-06add a cimple-valgrind executableEgor Tensin
2023-05-06add a TODO noteEgor Tensin
2023-05-06shut down server/workers gracefully on SIGTERMEgor Tensin
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-28 20:00:03 +0000