How to use the reader?

[Ambrevar]

> (in-package :iolib/pathnames)
> (iolib/base:enable-literal-reader* )
T
> #/p/"README.org"
#/P/"README.org"
> #/p/"README.org"
; error
> (iolib/pathnames:file-path "README.org")
#/IOLIB/PATHNAMES:P/"README.org"
> #/IOLIB/PATHNAMES:P/"README.org"
; error

Some things be wrong.

• The reader seems to only work in the IOlib package.
• Is it intended that the reader dispatch string changes?
• Why surround the reader dispatch string with /? Actually, just / would be a nice dispatch string for pathnames. For instance #/"path/to/foo" makes a lot of sense to me.

Thoughts?

[sionescu]

It works in any package that imports the symbol iolib/pathnames:p.

Is it intended that the reader dispatch string changes?

I'm not sure what you mean by this.

Why surround the reader dispatch string with /

I meant the #/ macro to be something generic that I could reuse. It currently works for IP addresses: #/ip/1.2.3.4 or #/ip/ff::1.

[Ambrevar]

It works in any package that imports the symbol iolib/pathnames:p.

OK, `(import 'iolib/pathnames:p)` did the trick. Maybe document it?

> Is it intended that the reader dispatch string changes? I'm not sure what you mean by this.

If you don't import the iolib/pathnames:p symbol, then the object is printed as above in `#/IOLIB/PATHNAMES:P/"README.org"`. This is not `cl:read`-able by the way. Bug then?

> Why surround the reader dispatch string with `/` I meant the `#/` macro to be something generic that I could reuse. It currently works for IP addresses: `#/ip/1.2.3.4` or `#/ip/ff::1`.

Understood. It's a bit verbose though. It could be convenient to have shorter forms. Say, `/` followed by double quotes would mean a pathname, while `/` not followed by double quotes would be an IP.