** Matches HTTP Requests to response objects.
**
** 'Route' is a mixin so you may provide your own implementations. The rest of this documentation
** relates to the default implementation which uses regular expressions to match against the
** Request URL and HTTP Method.
**
** Regex Routes
** ************
** Matches the HTTP Request URL and HTTP Method to a response object using regular expressions.
**
** Note that all URL matching is case-insensitive. If you really need case-sensitive matching (???)
** use the 'RegexRoute' explicitly, passing 'false' as the 'caseInsensitive' argument. Example:
**
** RegexRoute(`/index`, MyPage#hello, "GET", false)
**
**
**
** Response Objects
** ================
** A 'Route' may return *any* response object, be it `Text`, `HttpStatus`, 'File', or any other.
** It simply returns whatever is passed into the ctor.
**
** Example, this matches the URL '/greet' and returns the string 'Hello Mum!'
**
** Route(`/greet`, Text.fromPlain("Hello Mum!"))
**
** And this redirects any request for '/home' to '/greet'
**
** Route(`/home`, Redirect.movedTemporarily(`/greet`))
**
** You can use glob expressions in your URL, so:
**
** Route(`/greet.*`, ...)
**
** will match the URLs '/greet.html', '/greet.php' but not '/greet'.
**
**
**
** Response Methods
** ================
** Routes may also return `MethodCall` instances that call a Fantom method.
** To use, pass in the method as the response object.
** On a successful match, the 'Route' will convert the method into a 'MethodCall' object.
**
** Route(`/greet`, MyPage#hello)
**
** Method matching can also map URL path segments to method parameters and is a 2 stage process:
**
** Stage 1 - URL Matching
** ----------------------
** First a special *glob* syntax is used to capture string sections from the request URL.
** In stage 2 these strings are used as potential method arguments.
**
** In brief, the special glob syntax is:
** - '?' optionally matches the last character,
** - '/*' captures a path segment,
** - '/**' captures all remaining path segments,
** - '/***' captures the remaining URL.
**
** Full examples follow:
**
** glob pattern URL captures
** --------------------------------------------
** /user/* --> /user/ => ""
** /user/* --> /user/42 => "42"
** /user/* --> /user/42/ => no match
** /user/* --> /user/42/dee => no match
**
** /user/*/* --> /user/ => no match
** /user/*/* --> /user/42 => no match
** /user/*/* --> /user/42/ => "42", ""
** /user/*/* --> /user/42/dee => "42", "dee"
**
** /user/** --> /user/ => ""
** /user/** --> /user/42 => "42"
** /user/** --> /user/42/ => "42", ""
** /user/** --> /user/42/dee => "42", "dee"
**
** /user/*** --> /user/ => ""
** /user/*** --> /user/42 => "42"
** /user/*** --> /user/42/ => "42/"
** /user/*** --> /user/42/dee => "42/dee"
**
** Note that in stage 2 empty strings may be converted to 'nulls'.
**
** The intention of the '?' character is to optionally match a trailing slash. Example:
**
** glob URL
** -----------------------------
** /index/? --> /index => match
** /index/? --> /index/ => match
** vs
** /index/ --> /index => no match
** /index --> /index/ => no match
**
** Should a match be found, then the captured strings are further processed in stage 2.
**
** A 'no match' signifies just that.
**
**
**
** Stage 2 - Method Parameter Matching
** -----------------------------------
** An attempt is now made to match the captured strings to method parameters, taking into account nullable types
** and default values.
**
** First, the number of captured strings have to match the number of method parameters, taking into
** account any optional / default values on the method.
**
** Next the captured strings are converted to method arguments using the [ValueEncoder]`ValueEncoder`
** service. If no value encoder is found then the following default behaviour is used:
** - Non-empty strings are converted using a [TypeCoercer]`afBeanUtils::TypeCoercer`
** - Empty strings are considered 'null' but
** - if the method parameter is not nullable, then [BeanFactory.defaultValue()]`afBeanUtils::BeanFactory.defaultValue` is used.
**
** The above process may sound complicated but in practice it just works and does what you expect.
**
** Here are a couple of examples:
**
** strings method signature args
** ---------------------------------------------------------------
** --> (Obj a, Obj b) => no match
**
** "" --> (Str? a) => null
** "" --> (Str a) => ""
** "wotever" --> (Str a) => "wotever"
**
** "" --> (Int? a) => null
** "" --> (Int a) => 0
** "68" --> (Int a) => 68
** "wotever" --> (Int a) => no match
**
** "" --> (Str? a, Int b := 68) => null, (default)
** "" --> (Str a, Int b := 68) => "", (default)
**
** Assuming you you have an entity object, such as 'User', with an ID field; you can contribute a
** 'ValueEncoder' that inflates (or otherwise reads from a database) 'User' objects from a string
** version of the ID. Then your methods can declare 'User' as a parameter and BedSheet will
** convert the captured strings to User objects for you!
**
**
**
** Method Invocation
** -----------------
** Handler methods may be non-static.
** They they belong to an IoC service then the service is obtained from the IoC registry.
** Otherwise the containing class is [autobuilt]`afIoc::Registry.autobuild`.
** If the class is 'const', the instance is cached for future use.
**
const mixin Route {
@NoDoc @Deprecated { msg="Deprecated with no replacement. As Route is now a mixin, implementations matches may not be based on a regex." }
virtual Regex routeRegex() { Str.defVal.toRegex }
@NoDoc @Deprecated { msg="Deprecated with no replacement. As Route is now a mixin, implementations matches may not be based on HTTP methods." }
virtual Str httpMethod() { Str.defVal }
@NoDoc @Deprecated { msg="Deprecated with no replacement. As Route is now a mixin, implementations may generate dynamic responses." }
virtual Obj response() { Str.defVal }
** Creates a Route that matches on the given URL glob pattern.
** 'urlGlob' must start with a slash "/". Example:
**
** Route(`/index/**`)
**
** Note that matching is made against URI patterns in [Fantom standard form]`sys::Uri`.
** That means certain delimiter characters in the path section will be escaped with a
** backslash. Notably the ':/?#[]@\' characters. Glob expressions have to take account
** of this.
**
** 'httpMethod' may specify multiple HTTP methods, separated by spaces and / or commas.
** Each may also be a glob pattern. Example, all the following are valid:
** - 'GET'
** - 'GET HEAD'
** - 'GET, HEAD'
** - 'GET, H*'
**
** Use the simple string '*' to match all HTTP methods.
static new makeFromGlob(Uri urlGlob, Obj response, Str httpMethod := "GET") {
RegexRoute(urlGlob, response, httpMethod)
}
** For hardcore users; make a Route from a regex. Capture groups are used to match arguments.
** Example:
**
** Route(Regex<|(?i)^\/index\/(.*?)$|>, #foo, "GET", true) ==> Route(`/index/**`)
**
** Set 'matchAllSegs' to 'true' to have the last capture group mimic the glob '**' operator,
** splitting on "/" to match all remaining segments.
**
** Note that matching is made against URI patterns in [Fantom standard form]`sys::Uri`.
** That means certain delimiter characters in the path section will be escaped with a
** backslash. Notably the ':/?#[]@\' characters. Regular expressions have to take account
** of this.
**
** 'httpMethod' may specify multiple HTTP methods, separated by spaces and / or commas.
** Each may also be a glob pattern. Example, all the following are valid:
** - 'GET'
** - 'GET HEAD'
** - 'GET, HEAD'
** - 'GET, H*'
**
** Use the simple string '*' to match all HTTP methods.
static new makeFromRegex(Regex uriRegex, Obj response, Str httpMethod := "GET", Bool matchAllSegs := false) {
RegexRoute(uriRegex, response, httpMethod, matchAllSegs)
}
** Returns a response object should the given uri (and http method) match this route. Returns 'null' if not.
abstract Obj? match(HttpRequest httpRequest)
** Returns a hint at what this route matches on. Used for debugging and in 404 / 500 error pages.
abstract Str matchHint()
** Returns a hint at what response this route returns. Used for debugging and in 404 / 500 error pages.
abstract Str responseHint()
}