massimocandela / geofeed-finder

Utility to find geofeed files linked from rpsl.
BSD 3-Clause "New" or "Revised" License
74 stars 8 forks source link

Attempt to fixup Github URLs #35

Closed andrewachen closed 5 months ago

andrewachen commented 5 months ago

There are currently 39 geofeeds that are hosted on Github, but have the URL incorrectly set on their prefixes. This attempts to fixup these URLs so that those prefixes can be ingested. While the ideal solution would be for network owners to fix this themselves, this feels like a fairly easy fix that covers multiple network owners.

andrewachen commented 5 months ago

Thanks for reviewing this. Do you have suggestions for alternative names for this method?

I noticed that there were messages when HTML was encountered, so I thought that outputting logs when transforming the URL would be ok. Would you prefer that I remove the log message altogether?

massimocandela commented 5 months ago

Dear @andrewachen,

I really appreciate code contributions. However, I'm not going to merge this PR. This tool has two main goals: 1) Provide a working implementation of RFC9092; 2) Provide a tool to tests geofeed setup.

By tampering with URLs to overcome operators' setup errors, I would be compromising both objectives (including breaking the validation process). Additionally, I would be further encouraging misuse. If operators direct whois entries to HTML pages instead of CSV files, contrary to what is prescribed by the RFC, it's likely they are not so interested in resolving their geolocation issues.

andrewachen commented 5 months ago

@massimocandela that's very fair feedback, thank you. I noticed that there are command line arguments -u and -k that retain geofeeds that are technically invalid. Would this change be more acceptable if it were opt-in behind a flag?