.Dd $Mdocdate: September 10 2023 $
.Dt GSP 5
.Os
.Sh NAME
.Nm gsp
.Nd language reference for gsp
.Sh DESCRIPTION
The
.Nm
language is an alternative language to HTML and XML which can be transpiled into
either by making use of the
.Xr gsp 1
transpiler.
The
.Nm
language allows you to structure data in the same manner as XML while offering
an easier\-to\-read and less verbose syntax, and also some nice shortcuts for
working with HTML.  An example
.Nm
document might look as follows:
.Bd -literal -offset indent
!html{}

html lang="en" {
  head {
    meta charset="UTF-8"
    meta name="viewport" content="width=device-width, initial-scale=1.0"
    link href="/favicon.svg" rel="shortcut icon" type="image/svg"
    link href="/style.svg" rel="stylesheet"
    title {-My Website Title}
  }
  body {
    p #my-id  {- This is a paragraph with the id ‘my-id’     }
    p .my-cls {- This is a paragraph with the class ‘my-cls’ }

    p
      #some-id
      .class-1
      .class-2
      key-1="value-1"
      key-2 = "value-2"
    {-
      This paragraph has an ID, two classes, and two additional attributes.  GSP
      allows us to use the ‘#ident’ and ‘.ident’ syntaxes as shorthands for
      applying IDs, and classes.  This is a text node, so nothing is being
      interpreted as GSP nodes, but we can include them inline if we want.  As
      an example, here is some @em {-emphatic} text.  Your inline nodes can also
      have attributes @em #id .cls {-just like a regular node}.
    }
  }
}
.Ed
.Pp
As can be seen in the above example, nodes in
.Nm
take the form of
.Ql name attributes {...}
as opposed to the more verbose
.Ql <name attributes>...</name> .
.Sh LANGUAGE SYNTAX
.Ss Nodes
Nodes are the central building block of a
.Nm
document, and take the form of
.Ql name attributes {...} .
For example, a footer node with the attributes
.Sq foo
and
.Sq bar
and no child nodes is written as:
.Pp
.Bd -literal -offset indent
footer foo bar {}
.Ed
.Pp
To nest nodes within other nodes, simply place them within the braces.
As such, the following defines a footer with the attributes
.Sq foo
and
.Sq bar
with two empty div\-nodes as children:
.Pp
.Bd -literal -offset indent
footer foo bar { div{} div{} }
.Ed
.Ss Node names
Node names follow the exact same naming rules as names do in XML.
See the XML references in
.Sx SEE ALSO
for more details.
.Ss Attributes
Attributes are optional components of a node.
They take the form of an attribute name and an optional attribute value.
To specify an attribute, simply write the attribute name:
.Pp
.Bd -literal -offset indent
name
.Ed
.Pp
If you want to provide a value, you must follow the name with an equals sign
.Pq Sq =
and then wrap the value in double quotes
.Pq Sq \(dq :
.Pp
.Bd -literal -offset indent
name="value"
.Ed
.Pp
You can optionally include whitespace for visual clarity, and double quotes and
backslashes
.Pq Sq \e
can be escaped using a backslash:
.Pp
.Bd -literal -offset indent
name = "he said \e"hello there\e""
.Ed
.Pp
Like with node names, the details about which characters are allowed within an
attribute name are detailed in the XML reference found in the
.Sx SEE ALSO
section of this manual.
.Ss HTML IDs and -classes
When transpiling to HTML, you will be wanting to use IDs and classes all of the
time.
Due to the frequency of use of these parts of HTML,
.Nm
offers a shorthand syntax for specifying them.
The shorthand syntax looks like the equivalent CSS selectors for said IDs and
classes.
Valueless attributes prefixed with a period
.Pq Sq \&.
or hash
.Pq Sq #
are transpiled to classes and IDs respectively.
Therefore the following two examples are identical:
.Bd -literal -offset indent
div #foo .bar .baz {
  div .bar {}
}
.Ed
.Bd -literal -offset indent
div id="foo" class="bar baz" {
  div class="bar" {}
}
.Ed
.Ss Document types
Both HTML and XML make use of document types at the start of the document.  In
HTML the document type is of the form
.Ql <!DOCTYPE attributes>
while in XML the document type is of the form
.Ql <?xml attributes?> .
You can specify a document type in
.Nm
by using the special node name
.Sq \&!
or
.Sq \&?
for an HTML- or XML document type respectively.
Unlike with the usual node name, you do not need to include whitespace between
the special node name and the attributes.
The following example includes a document type that will transpile into
.Ql <!DOCTYPE html> :
.Bd -literal -offset indent
!html{}
html {
  head {...}
  body {...}
}
.Ed
.Pp
Here is an example of using an XML document type:
.Bd -literal -offset indent
?version="1.1" encoding="UTF-8"{}
document {...}
.Ed
.Pp
You may only specify one document type, and it must be the first node of the
document.
.Ss Literal text
If you want to include literal text within a node you can make use of the
special node name
.Sq - .
Unlike with the usual node name, you do not need to include whitespace between
the special node name and the attributes.
The following example shows how you can set a page title and paragraph text:
.Bd -literal -offset indent
!html{}

html {
  head {
    title {-My Amazing Website}
  }

  body {
    p {-
      Welcome to my website!  Here on my website you can find cute cat pictures,
      amongst other things.
    }
  }
}
.Ed
.Pp
When writing literal text, all occurrences of
.Sq } ,
.Sq @ ,
and
.Sq \e
must be backslash escaped as they have special meaning.
.Ss Embedded nodes
If you want to embed a node within literal text, you can make use of an embedded
node.
Embedded nodes are exactly the same as regular nodes, but they are prefixed with
the at
.Pq Sq @
symbol.
For example if you want to emphasize some text in a paragraph, you could do the
following:
.Bd -literal -offset indent
p {-
  This is some text, but @em .my-class {-some} of it is emphatic!
}
.Ed
.Ss Whitespace control
The typical behavior of
.Nm
is to compact whitespace.
Here you can see a before\-and\-after of transpilation:
.Bd -literal -offset indent
Before

foo {
  bar{- Hello World }
  baz{-Hello World}
}
.Ed
.Bd -literal -offset indent
After

<foo><bar>Hello World</bar><baz>Hello World</baz></foo>
.Ed
.Pp
The one exception to this use is when using embedded nodes.
If your literal text contains an embedded node, then whitespace around the node
is preserved:
.Bd -literal -offset indent
Before

foo {-
  Hello @bar{-there} world!
}
.Ed
.Bd -literal -offset indent
After

<foo>Hello <bar>there</bar> world!</foo>
.Ed
.Pp
Therefore if you would like to remove the whitespace, you need to manually
compact your document:
.Bd -literal -offset indent
Before

foo {-
  Hello@bar{-there}world!
}
.Ed
.Bd -literal -offset indent
After

<foo>Hello<bar>there</bar>world!</foo>
.Ed
.Sh SEE ALSO
.Xr gsp 1
.Pp
.Lk https://www.w3.org/TR/xml "Extensible Markup Language (XML) Reference"
.Sh AUTHORS
.An Thomas Voss Aq Mt mail@thomasvoss.com