This package provides a REST-capable, entirely non-blocking HTTP client library.
rl_http instvar varname METHOD url ?-option value ...?
| Option | Default | Description |
|---|---|---|
| -timeout | 15.0 | Time in seconds after which to consider the request a timeout. The timeout applies from the start of the connection attempt until the response is fully received. Use a value of "" to disable |
| -ver | 1.1 | The HTTP version to declare in the request |
| -accept | */* | The Accept header to send with the request |
| -headers | The request headers to send, as a list similar to a dictionary but allowing duplicate keys: HTTP headers can be multivalued | |
| -sizelimit | If set, and the returned Content-Length is larger than this value, and exception will be raised: {RL HTTP READ_BODY TOO_BIG $content_length} | |
| -data | The body of the request. Must already be encoded to bytes | |
| -data_cb | If set, the value is used as a command prefix to invoke to write the request body to the socket. The socket channel is appended as the first argument. The channel is in binary mode for writing | |
| -data_len | If -data_cb is used, the -data_len option can be used to supply a Content-Length header in the request | |
| -override_host | If set, use the supplied value as the request Host header, otherwise default to the authority section of the supplied url | |
| -tapchan | If set, a stacked channel will be layered on top of the socket, with the -tapchan value used as the command prefix for the reflected channel handler. An example handler is provided as ::rl_http::tapchan, which logs the read and write events and the base64 encoded bytes on the wire, for debugging. Redefine ::rl_http::log to suit your environment (default writes to stderr) | |
| -useragent | Ruby Lane HTTP client | The value to send as the User-Agent header in the request |
| -max_keepalive_age | -1 | If >= 0, the maximum age of a keepalive connection |
| -max_keepalive_count | -1 | If >=0, the maximum number of requests on a keepalive connection |
| Method | Arguments | Description |
|---|---|---|
| code | The HTTP status code of the response | |
| headers | The response headers, as a dictionary with the headers as keys, normalized to lower case, and the values as a list (HTTP headers are multi-valued) | |
| body | The response body, decoded and interpreted in the charset according to the response headers |
rl_http uses gc_class, so instance management is best left to bound instance variables:
rl\_http instvar h GET https://raw.githubusercontent.com/RubyLane/rl\_http/master/README.md
switch -glob -- [$h code] {
2* {
puts "Got result:\n[$h body]"
puts "Headers: [$h headers]"
}
default {
puts "Something went wrong: [$h code]\n[$h body]"
}
}When $h is unset (usually because it went out of scope), or its value is changed, the instance of rl_http will be destroyed.
Response headers (returned by the headers method) are represented as a dictionary with the header names as the keys, normalized to lowercase. The values are a list (HTTP headers can be multi-valued)
Request body data supplied in the -data option must be fully encoded, matching the Content-Type request header. For text types this usually means utf-8, for images it should be the raw bytes of the image.
set json\_body {
{
"hello": "server",
"foo": 1234
}
}
# utf-8 is the default for application/json, could also be explicit: "application/json; charset=utf-8"
rl\_http instvar h PUT $url -headers {Content-Type application/json} -data [encoding convertto utf-8 $json\_body]set h [open avatar.jpg rb]
try {set image\_bytes [read $h]} finally {close $h}
rl\_http instvar h PUT $url -headers {Content-Type image/jpeg} -data $image\_bytesThis package is licensed under the same terms as the Tcl core.