Possible issue with thread management (Helgrind) #1476
Description
Git commit reference
d46bd98 (V10.1.3)
Describe the bug
Running the basic ping/pong bot example through helgrind (valgrind --tool=helgrind) creates a gigantic log of data races and mutex/rwlock issues.
To Reproduce
Steps to reproduce the behavior:
- Compile the base program shown at https://dpp.dev/firstbot.html, without
bot.on_log(). (I personally used the flags-ldpp -std=c++23 -g3) - Run
valgrind --tool=helgrind ./a.out - Observe!
Expected behavior
Helgrind should not be outputting so many "possible data race", "write lock granted on mutex/rwlock which is currently wr-held by a different thread", and so on.
Screenshots
Not applicable, however I can attach both a log file as well as the sample code (without the token heh) that produced this error.
System Details:
- OS: Arch Linux
- Discord Client used for testing: N/A
Additional context
This seems to happen regardless of the OS. I also attempted it on Alpine Linux and i could recreate it as well.
Oddly enough, -fsanitize=thread does not complain at all.
Here's the program that I used to recreate this issue:
#include <dpp/dpp.h>
const std::string BOT_TOKEN = "insert token here";
int main() {
dpp::cluster bot(BOT_TOKEN);
bot.on_slashcommand([&bot](const dpp::slashcommand_t& event) {
if (event.command.get_command_name() == "stop") {
event.reply("Pong!");
bot.shutdown(); // shutting the bot down also seems to throw a million issues!
}
});
bot.on_ready([&bot](const dpp::ready_t& event) {
(void)event;
if (dpp::run_once<struct register_bot_commands>()) {
bot.global_command_create(dpp::slashcommand("stop", "stop!", bot.me.id));
}
});
bot.start(dpp::st_wait);
}And the attached Valgrind log:
Sorry if this seems a bit hectic? I could not find another mention of helgrind in the issues, so I decided to open one myself.
Have a nice day!