How to use the reader? - Githubissues

Ambrevar commented 3 years ago

> (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 commented 3 years ago

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 commented 3 years ago

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.

sionescu / iolib

How to use the reader? #65