💾 Archived View for gemi.dev › gemini-mailing-list › 000566.gmi captured on 2023-11-04 at 12:56:30. Gemini links have been rewritten to link to archived content

View Raw

More Information

➡️ Next capture (2023-12-28)

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

[spec] adding a "magic number" for gemini files

cage <cage-dev (a) twistfold.it>

Hi!

I think  a lot of  guesswork is needed  (line finding link  lines with
gemini  scheme?), for  example, to  discriminate between  markdown and
gemini file format.

I wonder if  a magic number (three, maybe four  chars?) could be added
on the first line of each file.

Bye!
C.

Link to individual message.

Solderpunk <solderpunk (a) posteo.net>

On Sat Dec 26, 2020 at 4:32 PM CET, cage wrote:

> I think a lot of guesswork is needed (line finding link lines with
> gemini scheme?), for example, to discriminate between markdown and
> gemini file format.

When is this really necessary?  Both Gemini and HTTP provide a means to
inform the client what's coming down the pipe.  For files distributed
otherwise, filename conventions would seem to do the trick.

> I wonder if a magic number (three, maybe four chars?) could be added
> on the first line of each file.

This feels like too much complexity for too little benefit to me.
Plenty of people write their text/gemini content by hand with a standard
text editor.  Many of them will forget to add the magic number.  What
should a client do when it receives a file which the server declares as
text/gemini but which doesn't have the magic number?  This creates more
questions/problems than it solves.

Cheers,
Solderpunk

Link to individual message.

Petite Abeille <petite.abeille (a) gmail.com>



> On Dec 26, 2020, at 16:32, cage <cage-dev at twistfold.it> wrote:
> 
> I wonder if  a magic number (three, maybe four  chars?) could be added
> on the first line of each file.

No*. 

We still pay the price for that BOM abomination. No need to double down.

That said, there are other signs, such as link sequences '=> '.

Otherwise, text/gemini is, rightly, identified as text/plain;charset=utf-8;. 

Why do you need more?



Link to individual message.

cage <cage-dev (a) twistfold.it>

On Sat, Dec 26, 2020 at 04:34:25PM +0100, Solderpunk wrote:
> On Sat Dec 26, 2020 at 4:32 PM CET, cage wrote:
>
> > I think a lot of guesswork is needed (line finding link lines with
> > gemini scheme?), for example, to discriminate between markdown and
> > gemini file format.
>
> When is this really necessary?  Both Gemini and HTTP provide a means to
> inform the client what's coming down the pipe.  For files distributed
> otherwise, filename conventions would seem to do the trick.

I was  thinking abot file(1) and  libmagic in general.

> > I wonder if a magic number (three, maybe four chars?) could be added
> > on the first line of each file.
>
> This feels like too much complexity for too little benefit to me.
> Plenty of people write their text/gemini content by hand with a standard
> text editor.  Many of them will forget to add the magic number.  What
> should a client do when it receives a file which the server declares as
> text/gemini but which doesn't have the magic number?  This creates more
> questions/problems than it solves.

Anyway  i agree this was a silly proposal, thank you for you reply! :)
Bye!
C.

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great cage once stated:
> On Sat, Dec 26, 2020 at 04:34:25PM +0100, Solderpunk wrote:
> > On Sat Dec 26, 2020 at 4:32 PM CET, cage wrote:
> >
> > > I think a lot of guesswork is needed (line finding link lines with
> > > gemini scheme?), for example, to discriminate between markdown and
> > > gemini file format.
> >
> > When is this really necessary?  Both Gemini and HTTP provide a means to
> > inform the client what's coming down the pipe.  For files distributed
> > otherwise, filename conventions would seem to do the trick.
> 
> I was  thinking abot file(1) and  libmagic in general.

  I use libmagic for my server [1] and yet, I still use an extension to mark
a Gemini text file.

  -spc

[1]	https://github.com/spc476/GLV-1.12556

Link to individual message.

cage <cage-dev (a) twistfold.it>

On Sat, Dec 26, 2020 at 06:25:43PM -0500, Sean Conner wrote:

[...]

> > I was  thinking abo[u]t file(1) and  libmagic in general.
>
>   I use libmagic for my server [1] and yet, I still use an extension to mark
> a Gemini text file.

Do you mean you check the file extension for recognize a gemini file?
Or  you use  same kind  of heuristic  to prevent  libmagic to  returns
gemini file as text?

Bye!
C.

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great cage once stated:
> On Sat, Dec 26, 2020 at 06:25:43PM -0500, Sean Conner wrote:
> 
> [...]
> 
> > > I was  thinking abo[u]t file(1) and  libmagic in general.
> >
> >   I use libmagic for my server [1] and yet, I still use an extension to mark
> > a Gemini text file.
> 
> Do you mean you check the file extension for recognize a gemini file?
> Or  you use  same kind  of heuristic  to prevent  libmagic to  returns
> gemini file as text?

  In the configuration file, the user can specify a mapping of extensions to
MIME type, to which the default extension type that identifies a Gemini
index file is added (which for my server is ".gemini" but that can be
changed).

  During a request, the extension is checked against this mapping.  If (for
a non-Gemini file) this doesn't exist, only then is libmagic queried.

  -spc

Link to individual message.

cage <cage-dev (a) twistfold.it>

On Sun, Dec 27, 2020 at 05:18:08PM -0500, Sean Conner wrote:

[...]

>
>   In the configuration file, the user can specify a mapping of extensions to
> MIME type, to which the default extension type that identifies a Gemini
> index file is added (which for my server is ".gemini" but that can be
> changed).
>
>   During a request, the extension is checked against this mapping.  If (for
> a non-Gemini file) this doesn't exist, only then is libmagic queried.
>
>   -spc

Crystal clear, thank you! :)
C.

Link to individual message.

---

Previous Thread: [spec] <URL> is a UTF-8 erratum

Next Thread: [user] psa: killfile