💾 Archived View for gemi.dev › gemini-mailing-list › 000019.gmi captured on 2023-12-28 at 15:39:47. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2023-11-04)

🚧 View Differences

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

[HELP] A weird issue when downloading files

1. Julien Blanchard (julien (a) typed-hole.org)

Hello Geminauts,

I have a weird issue that came up when coding stuff on Asuka (my client) 
and maybe you can help. I?m adding support for opening whatever mime type 
a Gemini server is serving me.

The logic I implemented so far is:
- serve text/* in the client as usual
- serve the other mime types by downloading content to a temporary file 
and opening it with xdg-open, open...

In order to download the served content I first check the header as 
string, get its size as bytes, re-fetch the whole ?page? as bytes, reverse 
the bytes array, truncate it with the header length and finally reverse it 
again before writing it to a file.

My issue is I have to add an offset of 5 bytes to the truncation for the 
files to be readable and I have no idea why. I tested this offset on 
several servers that are serving images, tarballs...  and it looks consistent.

Does it ring a bell by any chance? Or maybe I?m doing all of this completely wrong.

Link to individual message.

2. Sean Conner (sean (a) conman.org)

It was thus said that the Great Julien Blanchard once stated:
> Hello Geminauts,
> 
> I have a weird issue that came up when coding stuff on Asuka (my client)
> and maybe you can help. I?m adding support for opening whatever mime type
> a Gemini server is serving me.
> 
> The logic I implemented so far is:
> - serve text/* in the client as usual
> - serve the other mime types by downloading content to a temporary file
>   and opening it with xdg-open, open...
> 
> In order to download the served content I first check the header as
> string, get its size as bytes, re-fetch the whole ?page? as bytes, reverse
> the bytes array, truncate it with the header length and finally reverse it
> again before writing it to a file.

  I'm not understanding what you are doing here.

  Here's how I would do this---read bytes until you have seen the CRLF [1].
That's the response from the server.  I then would parse that line of data
to get the response and MIME type.  Then all I need to do is continue
reading data from the server and write that data toa file.  Then launch the
external program.

> My issue is I have to add an offset of 5 bytes to the truncation for the
> files to be readable and I have no idea why. I tested this offset on
> several servers that are serving images, tarballs...  and it looks
> consistent.
> 
> Does it ring a bell by any chance? Or maybe I?m doing all of this
> completely wrong.

  Not really.  And whatever you are doing it seems overly complicated ... oh
wait ... you aren't using Python 3 by any chance, are you?

  -spc (just a stab in the dark ... )

[1]	ASCII character 13 and ASCII character 10

Link to individual message.

3. Julien Blanchard (julien (a) typed-hole.org)


On 10/2/19 12:04 AM, Sean Conner wrote:
>    I'm not understanding what you are doing here.
I wanted to remove the header part in the most convoluted way possible :D
>    Here's how I would do this---read bytes until you have seen the CRLF [1].
> That's the response from the server.  I then would parse that line of data
> to get the response and MIME type.  Then all I need to do is continue
> reading data from the server and write that data toa file.  Then launch the
> external program.

OMG you're so right! I was trying to put a square peg in a round hole 
since I already
had the content splitted has two strings (header/body) when I should 
have re-thought
my parsing strategy the way you're describing directly at the bytes level.

>    Not really.  And whatever you are doing it seems overly complicated ... oh
> wait ... you aren't using Python 3 by any chance, are you?
>
>    -spc (just a stab in the dark ... )
>
> [1]	ASCII character 13 and ASCII character 10
I'm using Rust but I'm curious what the issue could be with python 3?

Anyway thanks a lot Sean! It helped a lot!

Link to individual message.

---

Previous Thread: TLS, TOFU and the Gemini security model

Next Thread: gusmobile, a Python library for Gemini requests