I'm on 4.0.8. I don't know what you mean by adding file: at the
beginning. They are file uris so they already have file:// at the
beginning. It does look from the code at the bottom of this file
https://github.com/savonet/liquidsoap/blob/77b6406f094b0732b41832ebaa5e2d33ab7218e4/libs/protocols.liq
that it's doing just what I thought and treating it like it's just a
regular path (not url decoding). And just to clarify, url decode would
mean converting %20 to space.
On Sun, May 17, 2020 at 6:13 PM Romain Beauxis <romain.beau...@gmail.com> wrote:
>
> Hi Thomas,
>
> That's an annoying issue indeed.
>
> What version of liquidsoap are you using? If your version is recent enough,
> you might be able to add a `file:` protocol that would take care of this for
> you.
>
> Matter of fact, I've also added it to our latest `master` branch.. You should
> be able to just add it to your script:
>
> def file_protocol(~rlog,~maxtime,arg) =
> path = list.nth(default="", string.split(separator=":",arg), 1)
> segments = string.split(separator="/",path)
>
> segments =
> if list.length(segments) > 3 and
> list.nth(default="foo",segments,0) == "" and
> list.nth(default="foo",segments,1) == ""
> then
> list.tl(list.tl(segments))
> else
> segments
> end
>
> path = string.concat(separator="/",segments)
>
> # We support the buggy form: file://path/to/file"
> if file.exists("/#{path}") then
> ["/#{path}"]
> else
> [path]
> end
> end
>
> add_protocol(static=true, temporary=false, "file", file_protocol,
> doc="File protocol. Only local files are supported",
> syntax="file:///path/to/file")
>
>
> Le sam. 16 mai 2020 à 12:59, Thomas J. Webb <tho...@thomaswebb.net> a écrit :
>>
>> I just started using liquidsoap and it's awesome and solving a lot of
>> problems for me. One minor issue I did have though is for when I tried
>> playing a playlist I exported from audacious, it gets confused by
>> files containing spaces. Audacious uses file:// uris instead of file
>> paths and the uris are valid. It looks like when liquidsoap encounters
>> a file uri, simply chops off the file:// part, then treats it like a
>> regular file path instead of url decoding first:
>>
>> file:///mp3/PIRATESHIPS.mp3 <- works fine
>>
>> file:///mp3/taste%20in%20men.mp3 <- gives me "Nonexistent file or
>> ill-formed URI ..."
>>
>> I encounter code doing this a lot and I've probably even written code
>> that did this but unfortunately this heuristic causes problems.
>>
>>
>> _______________________________________________
>> Savonet-users mailing list
>> Savonet-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/savonet-users
>
> _______________________________________________
> Savonet-users mailing list
> Savonet-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/savonet-users
--
Thomas J. Webb
+1 760-912-1856 (US Cell)
handfetisch (Skype)
_______________________________________________
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
