aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/src (follow)
Commit message (Collapse)AuthorAge
* worker: actually stay offlineEgor Tensin2023-07-05
| | | | | I don't know what I was thinking, but contrary to my intention, the worker stayed connected to the server all the time.
* tcp_server: minor refactoringEgor Tensin2023-07-05
|
* tcp_server: close client connectionsEgor Tensin2023-07-05
| | | | | I kinda forgot that I'm supposed to clean up client connections, and they immediately blew up in my face.
* command: dedupe the codeEgor Tensin2023-07-05
|
* tcp_server: keep track of client threadsEgor Tensin2023-07-05
| | | | | | | This is a major change, obviously; brought to me by Valgrind, which noticed that we don't actually clean up after cimple-client threads. For a more thorough explanation, please see the added comment in tcp_server.c.
* sanitize #include-sEgor Tensin2023-07-04
|
* worker: close the leftover descriptorEgor Tensin2023-07-04
| | | | | Thanks, Valgrind! As a note: if I think that Valgrind reports a false positive, chances are, it's not.
* move custom message parsing to a separate moduleEgor Tensin2023-07-04
|
* storage_sqlite: refactoringEgor Tensin2023-07-04
|
* sqlite: rename a columnEgor Tensin2023-07-04
|
* storage: mark completed runs as suchEgor Tensin2023-07-04
|
* storage: requeue old runs from storage on startupEgor Tensin2023-07-04
|
* tcp_server: always clean up connection descriptorsEgor Tensin2023-07-04
|
* sqlite: store new runs in SQLiteEgor Tensin2023-07-04
|
* storage_sqlite: refactoringEgor Tensin2023-07-04
|
* sqlite: add run status fieldEgor Tensin2023-07-04
|
* fix a compiler warningEgor Tensin2023-07-04
|
* appease clang-formatEgor Tensin2023-07-04
|
* use designated struct initializers moreEgor Tensin2023-06-30
|
* show git hash with --versionEgor Tensin2023-06-30
| | | | Also, use cmake's configure_file to build string constants in.
* fix function names in error messagesEgor Tensin2023-06-30
|
* use __WNOTHREAD w/ waitpidEgor Tensin2023-06-30
| | | | | | | | | | I learned about this flag on my previous job; basically, it should be enabled by default IMO, so that the thread doesn't receive signals from other threads' children. Here, it doesn't matter too much, since we're waiting for a specific child. However, if we were to use waitpid(-1, ...), it would be essential to use this flag. Still, even here, it's good to have this on.
* log: minor refactoringEgor Tensin2023-06-28
|
* process: fix function name in logsEgor Tensin2023-06-28
|
* ci: environment array to execvpe must be NULL-terminatedEgor Tensin2023-06-28
| | | | Thanks, Valgrind.
* log: support logging levelsEgor Tensin2023-06-28
|
* test: skip ci.sh w/ ValgrindEgor Tensin2023-06-28
|
* valgrind: exit with an error after leak detectionEgor Tensin2023-06-28
| | | | I cannot believe this isn't default.
* file: readlink(2) doesn't null-terminateEgor Tensin2023-06-28
| | | | Thanks, valgrind.
* minor refactoringEgor Tensin2023-06-14
|
* cmake: fix grammar in option descriptionEgor Tensin2023-06-13
|
* cmake: why the hell would I need MSVC optionsEgor Tensin2023-06-13
|
* minor refactoringEgor Tensin2023-06-13
|
* signal: remove the stupid add_to_event_loop wrapperEgor Tensin2023-06-13
|
* server: handle disconnected workers gracefullyEgor Tensin2023-06-13
|
* event_loop: support more event typesEgor Tensin2023-06-13
|
* event_loop: add event_loop_add_onceEgor Tensin2023-06-13
|
* signal: sigprocmask -> pthread_sigmaskEgor Tensin2023-06-13
|
* signal: refactoringEgor Tensin2023-06-13
|
* sanitize #include-sEgor Tensin2023-06-13
|
* signal: remove unused APIsEgor Tensin2023-06-13
|
* event_loop: hide the API detailsEgor Tensin2023-06-13
|
* use signalfd to stop on SIGTERMEgor Tensin2023-06-13
| | | | | | | | | | | | | | | | | | 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.
* log: refactoringEgor Tensin2023-06-12
|
* log: thread-safe loggingEgor Tensin2023-06-12
|
* log: prefix with thread IDEgor Tensin2023-06-12
|
* command: fix a race conditionEgor Tensin2023-06-12
| | | | | Modifying cmd_dispatcher fields like that make it inherently unsafe to call cmd_dispatcher_handle_conn concurrently.
* sqlite: fix generate-sql-header.pyEgor Tensin2023-06-12
|
* msg: rework some APIsEgor Tensin2023-06-11
|
* worker_queue: fix worker_queue_add_lastEgor Tensin2023-06-09
|
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-25 04:02:16 +0000