jesopo
jesopo
A good note about this from #142: > In addition to the Tor address obviously needing to be configurable, I think Bitbot should also support SOCKS proxy authentication, maybe with...
the proxy configuration system would be able to utilise http, https, socks4 and socks5 proxies at a minimum :)
Might just bite the bullet on this one and have a module that overwrites `utils.http.request()` in its `on_load()` and rewrites it in `unload()` that'll at least solve the problem for...
love this idea btw
the only issue I have with this is that the process to "get latest toot" is 4 to 7 HTTP requests - and that can take a long time. I...
deadline for slow transfers https://github.com/jesopo/bitbot/blob/4a6037c77405f3584efadc10ae75826b6b9ac422/src/utils/http.py#L240 max file size for large files that could oom the bot https://github.com/jesopo/bitbot/blob/4a6037c77405f3584efadc10ae75826b6b9ac422/src/utils/http.py#L224 don't know what problem you've managed to stumble on, but there's already code...
I'm sure the machine it's running on can read and parse 100mib of html unless it's somehow akin to a zipbomb. can you get the stacktrace from ben? we're going...
I've hit it with much worse in testing. eager to see a stacktrace
if it is a timeout, especially on the `.soup()` call outside the deadline, I'd be inclined to do a much less thorough parse, even just a regex to grab the...
benchmarking lxml against html5lib puts the former far ahead of the latter but I recall picking the latter for fault tolerances the former doesn't have