Print HTML title from untrusted URL

urltitle.d

import core.time, std.algorithm, std.range,
	   std.string, std.net.curl, std.regex;

import std.stdio : stderr, stdout;

immutable usage = `%s <URL to HTML page>`;

enum StatusCode
{
	success = 0,
	missingURL,
	unsupportedProtocol,
	timeout,
	missingTitle,
}

int main(string[] args)
{
	if(args.length < 2)
	{
		stderr.writefln(usage, args[0]);
		return StatusCode.missingURL;
	}

	auto url = args[1];
	static urlPattern = ctRegex!`^([a-z]+)://`;

	if(auto m = url.matchFirst(urlPattern))
	{
		auto prefix = m.captures.front;
		auto protocol = m.captures.drop(1).front;
		if(protocol != "http")
		{
			stderr.writefln("unsupported protocol: %s", protocol);
			return StatusCode.unsupportedProtocol;
		}
		url.skipOver(prefix);
	}

	auto http = HTTP(url);
	http.method = HTTP.Method.get;
	http.operationTimeout = 5.seconds;
	http.maxRedirects = 5;

	string contentType;
	string charset;

	http.onReceiveHeader = (key, value) {
		if(key == "content-type")
		{
			// parse "text/html; charset=xxx"
			auto fields = value.splitter(";").map!(field => field.strip);
			contentType = fields.front.idup;

			auto charsetParam = fields.drop(1)
				.find!(field => field.startsWith("charset="));

			if(!charsetParam.empty)
			{
				charset = charsetParam
					.front
					.splitter("=")
					.drop(1)
					.front
					.idup;
			}
		}
	};

	enum sizeThreshold = 1024 * 8;
	ubyte[sizeThreshold] buffer;
	auto bufferTail = buffer[];

	http.onReceive = (data) {
		if(contentType && contentType != "text/html")
			return HTTP.requestAbort;

		if(bufferTail.length < data.length)
		{
			immutable partLength = bufferTail.length;
			bufferTail = data
				.take(partLength)
				.copy(bufferTail);
			return HTTP.requestAbort;
		}
		else
		{
			bufferTail = data.copy(bufferTail);
			return data.length;
		}
	};

	try http.perform();
	catch(CurlTimeoutException ex)
	{
		stderr.writeln(ex.msg);
		return StatusCode.timeout;
	}
	catch(CurlException)
	{
		// Assume aborted... :S
	}

	auto document = buffer[0 .. buffer.length - bufferTail.length];
	string transcoded;

	if(icmp(charset, "utf-8") != 0)
	{
		// TODO: transcode to UTF-8
		transcoded = cast(string)document;
		stderr.writefln(`WARNING: unrecognized charset "%s"`, charset);
	}
	else
		transcoded = cast(string)document;

	// HE COMES
	static titlePattern = ctRegex!(`<title>(.*?)</title>`, "si");

	if(auto m = transcoded.matchFirst(titlePattern))
		stdout.writeln(m.captures.drop(1).front);
	else
	{
		stderr.writefln("unable to find title near start of document (within %s bytes)", sizeThreshold);
		return StatusCode.missingTitle;
	}

	return StatusCode.success;
}