💾 Archived View for rawtext.club › ~sloum › geminilist › 004765.gmi captured on 2023-11-04 at 14:25:08. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2021-11-30)
-=-=-=-=-=-=-
Jon jon at shit.cx
Thu Jan 7 08:14:42 GMT 2021
- - - - - - - - - - - - - - - - - - -
On 07/01/21 02:25, Sean Conner wrote:
It was thus said that the Great Jon once stated:
On 06/01/21 22:00, Jason McBrayer wrote:
I think it's a good idea in principle, but it is maybe not
well-suited for multi-user sites, which a lot of Gemini sites are.
Hmm. I assumed that most sites were on their own domain.
On a multi-tenanted host, is each tenant generally confined to a
unix homedir? For example `/~${USER}/*`. If that's the case,
checking the users root before hitting the root shouldn't be
unreasonable.
It might not always be the case that a user's directory is marked by
a '~'. For instance, on <gemini://republic.cirumlunar.space/> users
are under <gemini://republic.circumlunar.space/users/>.
-spc
I guess traversing up the tree looking for a humans.txt file would work.
Finding the humans.txt file for a document at`/users/jon/gemlog/2021-01-07-a-new-doc/index.gmi` would search:
- /users/jon/gemlog/2021-01-07-a-new-doc/humans.txt- /users/jon/gemlog/humans.txt- /users/jon/humans.txt- /users/humans.txt- /humans.txt
This would allow one document, or an entire branch of the tree have adifferent contact than the rest, but it's a lot more complex than simplyhitting up /humans.txt.
Even though it handles the edge cases, I'm less enthusiastic about thissolution. What was initially a simple task of fetching /humans.txt is,well not that anymore.
-- Jon