Content-Disposition header?

[Mark Stosberg]

Providers usually don't include the `Content-Disposition` header, but today I ran across one that does. 

One difference with the Content-DIsposition header is that it includes the file name for the file that they want to use, versus having the file name in the URL.

The difference confused a tech here, but doesn't seem like it's going to gum up our automation. 

Yet, just to keep things simpler and consistent, it seems worth considering recommending against using this header. Alternately, if it's intended to be supported, it could be worth mentioning in the spec. 

[Elias Gino Cripotos]

Hello Mark, 

Thank you for your question! 

If I am not mistaken, your are talking about GTFS-Realtime correct? Would you have a realtime feed to share with the community? 

I would highly recommend to post this as an issue on Github or even in the GTFS-RT slack channel. You will get quicker responses from passionate community members. 

Let me know if you have any questions, 

Elias 

[Mark Stosberg]

If I am not mistaken, your are talking about GTFS-Realtime correct? Would you have a realtime feed to share with the community? 

This was about GTFS, not GTFS-RT. 

I ran across a feed in the wild that sent a Content-Disposition header, which seemed strange, and something that the spec might want to clarify. 

     Mark

[Stefan de Konink]

On Thursday, August 3, 2023 4:16:42 PM CEST, Mark Stosberg wrote:
> I ran across a feed in the wild that sent a Content-Disposition header,
> which seemed strange, and something that the spec might want to clarify.

If it is a symlink, to get a stable URI, pointing to the actual filename.
What is, in your opinion, wrong with that practice?

--
Stefan

[Mark Stosberg]

My take is that if Content-Disposition headers are intended to be supported then that should be documented in the spec.

If they are not intended to be supported,  perhaps the spec should explicitly ignore or forbid them because the parsing of the HTTP message body is a bit different for responses that use this header. 

      Mark

--
Stefan

[Stefan de Konink]

On Monday, August 7, 2023 6:56:55 PM CEST, Mark Stosberg wrote:
> My take is that if Content-Disposition headers are intended to be supported
> then that should be documented in the spec.
> If they are not intended to be supported, perhaps the spec should
> explicitly ignore or forbid them because the parsing of the HTTP message
> body is a bit different for responses that use this header.

This spec is about the content of the zip-delivery, not how it is
delivered.

--
Stefan

[Guillaume Campagna]

I agree with Stefan, I don’t think the zip file name matters at all when using a GTFS so I don’t see the benefit of adding this in the spec. 

—

Guillaume Campagna

CTO, Transit

transit.app

--
You received this message because you are subscribed to the Google Groups "GTFS Changes" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gtfs-changes...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gtfs-changes/866ec81c-c85e-42b7-8b49-a5aaedc39d7b%40konink.de.