Closed hugomrdias closed 8 months ago
So, this is the relevant code for this request:
It does indeed look for a DNS record at _did.<domain>
.
I think you may be misunderstanding something. E.g. I'd be issuing a request PATCH /api/v0/account/handle/matheus23.com
and have a DNS TXT record at _did.matheus23.com
set to did:key:<accountDID>
.
There's no more username, only the handle.
so the record is on _did.<domain>
not on _did.<username>.<domain>
? but the local doh is on _did.<username>.localhost
what if domain owner whats to do multiple accounts ?
on the same topic i wrote some notes on username/handle :
hugomrdias
hugomrdias.fission.codes
username
as <username>.<domain>
username
as <username>
handle
as just the domainwe should describe better username
, handle
and domain
in the context of this api
what if domain owner whats to do multiple accounts ?
Then the domain owner should/could use subdomains. I.e. work.matheus23.com
and personal.matheus23.com
.
I think that's better than requiring them to add an additional subdomain. What if I do want to use matheus23.com
? I don't want to be forced to use fission.matheus23.com
or matheus23.matheus23.com
, that'd be weird.
we should describe better
username
,handle
anddomain
in the context of this api
Yep. :100:% Agree. Should we just use username
as e.g. matheus23
, and handle
as matheus23.com
or matheus23.computer.name
? (or .localhost
if running locally. Eventually I think it'll be .computer.name
when deployed).
Also - what is your opinion on having handles & usernames simultaneously. I felt like it'd make sense to always have a username, since that's what we have control over. Otherwise it's possible that domain names expire, what would a user's username be then? That's why all of the things are somewhat weird right now. I only return the handle, if available in account info, not the username, even though in theory it still exists. But at the same time the patch handle & patch username routes are different. So it's weird.
Closing this in favor of #239
I think that's better than requiring them to add an additional subdomain. What if I do want to use
matheus23.com
? I don't want to be forced to usefission.matheus23.com
ormatheus23.matheus23.com
, that'd be weird.
How does a service know that it should use the top level domain for hugomrdias
and not hugomrdias.hugodias.dev and for hugodias
it's need to use subdomain because top level is taken by hugomrdias
?
we should describe better
username
,handle
anddomain
in the context of this apiYep. 💯% Agree. Should we just use
username
as e.g.matheus23
, andhandle
asmatheus23.com
ormatheus23.computer.name
? (or.localhost
if running locally. Eventually I think it'll be.computer.name
when deployed).Also - what is your opinion on having handles & usernames simultaneously. I felt like it'd make sense to always have a username, since that's what we have control over. Otherwise it's possible that domain names expire, what would a user's username be then? That's why all of the things are somewhat weird right now. I only return the handle, if available in account info, not the username, even though in theory it still exists. But at the same time the patch handle & patch username routes are different. So it's weird.
Imo we should only have handle, by default and as a fallback we provide
I think that's better than requiring them to add an additional subdomain. What if I do want to use
matheus23.com
? I don't want to be forced to usefission.matheus23.com
ormatheus23.matheus23.com
, that'd be weird.How does a service know that it should use the top level domain for
hugomrdias
and not hugomrdias.hugodias.dev and forhugodias
it's need to use subdomain because top level is taken byhugomrdias
?
The domain should be the source of truth.
In a log-in form, you type hugodias.dev
or hugodias.computer.name
, depending on what you use/have set up.
If what you have is a DID, then you use the account/info
capability & endpoint, and you'll get the canonical name for the account, that'd be either hugodias.dev
or hugodias.computer.name
.
I think that's better than requiring them to add an additional subdomain. What if I do want to use
matheus23.com
? I don't want to be forced to usefission.matheus23.com
ormatheus23.matheus23.com
, that'd be weird.How does a service know that it should use the top level domain for
hugomrdias
and not hugomrdias.hugodias.dev and forhugodias
it's need to use subdomain because top level is taken byhugomrdias
?The domain should be the source of truth. In a log-in form, you type
hugodias.dev
orhugodias.computer.name
, depending on what you use/have set up. If what you have is a DID, then you use theaccount/info
capability & endpoint, and you'll get the canonical name for the account, that'd be eitherhugodias.dev
orhugodias.computer.name
.
agree, user set ups whatever he wants tld, subdomain, fission fallback (
Summary
Problem
When changing the handle for an account, where the domain does not have the TXT record yet the api returns this error:
and it should be