On October 8 2017, I noticed that an old twitter bot of mine was gaining a lot of followers. I tweeted about it for a bit, and went to bed. I woke up to find that TheNextWeb turned my tweets into an article. Genbeta, a similar Spanish-language site, picked it up. Even via Google Translate into English, the Genbeta article seems better written to me.
Firstly, the script powering the bot does not actually use regular expressions. It is much too simple and poorly written to do so. Secondly, attempts at addressing the bot's persistent scunthorpe problem involve tradeoffs between missing some potentially worthwhile messages.
It is a mess of a script that glues together two public APIs. If you want to learn how to make your own clone, the API documentation will be much more instructive than the source for the bot. While I am sure that I would receive many wonderful improvements to the bot if I made it open source, I am not interested enough in those potential improvements to put in the necessary work to maintain the bot as an open source project. Also, I don't like the idea of the bot tweeting commit messages to its own source.
A common first instinct when someone learns about the bot is to push commits to GitHub with unfunny messages in the hopes that they will get tweeted by the bot. I prefer to discourage this behaviour where possible. This behaviour would be encouraged rather than discouraged if the bot tweeted a link too. The other angle is that the author of the commit may not find the extra attention to be desirable. I would rather allow them to remain anonymous.
Each tweet is a commit message (or an excerpt thereof) from GitHub. Even edgy teens use GitHub. I don't filter what the bot tweets. It is a dumb script that tweets dumb things. It runs all day, every day and it is designed to do so with minimal day-to-day (or even year-to-year) time investment from me.
No. The script does not perform any logging.