💾 Archived View for gemi.dev › gemini-mailing-list › 000471.gmi captured on 2023-12-28 at 15:47:45. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2023-11-04)

🚧 View Differences

-=-=-=-=-=-=-

A question regarding section 1.2 of the spec

1. Ali Fardan (raiz (a) stellarbound.space)

1.2:
> The path, query and fragment components are allowed and have no
> special meanings beyond those defined by the generic syntax.

We know that the query portion of URIs is reserved for response code 10
and 11, but fragment is never used anywhere in the protocol.

So both query and fragment are dealt with absolutely differently in the
protocol, but both are categorized under the same "allowed and have no
special meanings", may I ask in what context do they have no special
meanings? I'm confused a bit.

Also, what does "defined by the generic syntax" refer to? what syntax?

Link to individual message.

2. cbabcock (a) asciiking.com (cbabcock (a) asciiking.com)

New here, but that never stops me from jumping in...

November 15, 2020 1:34 AM, "Ali Fardan" <raiz at stellarbound.space> wrote:

> 1.2:
> 
>> The path, query and fragment components are allowed and have no
>> special meanings beyond those defined by the generic syntax.
> 
> We know that the query portion of URIs is reserved for response code 10
> and 11, but fragment is never used anywhere in the protocol.

Your use of "reserved" here probably reflects a misunderstanding. You can 
have a URI with a query indicator without getting a 1x response code 
first. The word "reserved" in the spec only ever refers to characters that 
must be escaped.

> So both query and fragment are dealt with absolutely differently in the
> protocol, but both are categorized under the same "allowed and have no
> special meanings", may I ask in what context do they have no special
> meanings? I'm confused a bit.

A URI is best thought of as an opaque identifier. If it happens to have 
semantic or functional purpose for the server that mechanisms can leverage 
to improve the reader experience, great, but it's just a way to identify 
the desired content for delivery. There's a feedback mechanism to refine a 
query, but no corresponding mechanism for fragments. If there was, it 
would be something like the anchor tag in HTML.

> Also, what does "defined by the generic syntax" refer to? what syntax?

The generic URI syntax as defined in RFC 3986.

Chris

Link to individual message.

3. Ali Fardan (raiz (a) stellarbound.space)

On Sun, 15 Nov 2020 09:13:54 +0000
cbabcock at asciiking.com wrote:
> Your use of "reserved" here probably reflects a misunderstanding. You
> can have a URI with a query indicator without getting a 1x response
> code first.

To clarify, I was referring to being reserved for the use in a
standardized protocol specific function.

> The word "reserved" in the spec only ever refers to characters that
> must be escaped.

The word "reserved" refers to characters that must be escaped in RFC
3986, not in the Gemini spec, I was specifically referring to the case
where the query portion of the URI in 1X requests is reserved
specifically for user input therefore it can't be used for other
purpose in that specific case.

I was pointing out the fact that the Gemini spec referred to the query
portion of URI as "allowed and have no special meanings beyond those
defined by the generic syntax", this is confusing because A) the query
portion does have meaning in the context of user input responses, and in
that case it is actually reserved and can not be used for other purpose
B) "the generic syntax" is vague term.

What were you referring to by "the spec", did you mean RFC 3986 or the
Gemini spec?

Link to individual message.

---

Previous Thread: [ANN] -- kinda? LEO, Molniya, and the greater question of Gemini webrings

Next Thread: [ANN] Aquarius, a PHP app framework for Gemini CGI