Showing content from https://github.com/ipfs/go-ipfs/issues/4143 below:
Make base32 CIDv1 the default for go-ipfs · Issue #4143 · ipfs/kubo · GitHub
I understand there's a switch to CIDv1 soon. I think go-ipfs should use lowercased base32 (rfc4648 - no padding - highest letter) as the default multibase.
The reason this encoding is preferable: it's the one encoding that will work with subdomains (RFC1035 + RFC1123). The restrictions are: case-insensitive, a-b0-9 and less than 63 bytes.
For a slight increase in length, you reap enormous benefits:
- The ability to do proper security origins for the HTTP gateway with subdomains (cidv1abcde.dweb.link). This is very important if we want to handle reports with Google's safe browsing system (which is designed for origins). With the current design, all content is on the same browser origin, and a single phishing/malware report on any of the IPFS gateways (hosted by us or someone else) will make web browsers block every single thing on the origin with a giant red warning message until it's cleared up with Google (which from experience can take several days!)
- Root paths are in the right place, which dramatically improves compatibility with existing web sites that tend to do a lot of this:
<img src="/rootimg.jpg">
- Allows us to register dweb.link (and ipfs.io, etc.) to the Public Suffix List, which will prevent the sandboxed content from reading/manipulating cookies on the parent domain (and on other cidv1 subdomains).
- Opens up the ability for go-ipfs to do HTTP Host Header parsing and automatic Let's Encrypt support (if we wanted to), so anyone can set up a public IPFS gateway without additional software. Once Let's Encrypt gets their wildcard cert domains shipped (Dec 2017), this could be a fully automated process. Otherwise something like nginx would be needed (I could write an example nginx.conf that people could use for it).
It should use lowercase base32 characters by default, so that it's consistent with subdomain usage (all the browsers will force lowercase). IIRC the RFC doesn't care if it's lowercased, I think people just default to upper case for legacy reasons.
Obviously an abstraction layer could be written that converts between base32 and something else for use with web gateways and then have a different default, but I think it would less confusing for end users to use one default: the one that will let origins in browsers work.
This approach shouldn't be a problem for webextension plugins, but @lidel feel free to chime in.
Further reading: https://github.com/neocities/hshca
lidel, jdanford, jessebmiller, Leonix, burdakovd and 12 more
RetroSearch is an open source project built by @garambo
| Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4