/// <summary>
/// Take a list of strings like 'ftp', 'http', 'attachment' and append them as a full
/// searchable instance to the collection of schemes to find in the input, like
/// 'ftp://', 'ftp%3a//', 'http://', 'http%3a//' etc
/// </summary>
/// <param name="validSchemes"></param>
private void SetValidSchemes(IEnumerable <string> validSchemes)
{
ValidSchemesNames.Clear();
ValidSchemesSuffixed.Clear();
foreach (var validScheme in validSchemes)
{
var lowerInvariant = validScheme.Trim().ToLowerInvariant();
ValidSchemesNames.Add(lowerInvariant);
ValidSchemesSuffixed.Add(lowerInvariant + "://");
ValidSchemesSuffixed.Add(lowerInvariant + "%3a//");
}
}
/// <summary>
/// Reads the scheme and allows returns true if the scheme is in our allowed collection (e.g. http(s?):// or ftp(s?)://)
/// @return True if the scheme was found, else false.
/// </summary>
private bool ReadScheme()
{
var originalLength = _buffer.Length;
var numSlashes = 0;
while (!_reader.Eof())
{
var curr = _reader.Read();
//if we match a slash, look for a second one.
if (curr == '/')
{
_buffer.Append(curr);
if (numSlashes == 1)
{
// return only if the buffer currently ends with an approved protocol.
// When we have buffered a string like: ":u(https://test.co" and are scanning the ':', we
// consider this to have found a scheme (the https bit only, which will be parsed out later)
var bufferedUrlContent = _buffer.ToString().ToLowerInvariant();
// return only if we detect an approved protocol at the end of the current buffer. For
// efficiency, first check an exact match
if (ValidSchemesSuffixed.Contains(bufferedUrlContent))
{
_currentUrlMarker.SetIndex(UrlPart.SCHEME, 0);
return(true);
}
// If no exact match found, try to find a valid scheme in the trailing content of the current buffer,
// starting with the longest matches available (e.g. sftp:// rather than ftp://
for (var i = ValidSchemesSuffixedOrdered.Length - 1; i >= 0; i--)
{
var vss = ValidSchemesSuffixedOrdered[i];
if (bufferedUrlContent.EndsWith(vss))
{
// see if we need to remove extra characters from the start of the buffer
if (bufferedUrlContent.Length > vss.Length)
{
_buffer.Remove(0, bufferedUrlContent.Length - vss.Length);
}
_currentUrlMarker.SetIndex(UrlPart.SCHEME, 0);
return(true);
}
}
return(false);
}
numSlashes++;
}
else if (curr == ' ' || CheckMatchingCharacter(curr) != CharacterMatch.CharacterNotMatched)
{
//if we find a space or end of input, then nothing found.
_buffer.Append(curr);
return(false);
}
else if (curr == '[')
{
//if we're starting to see an ipv6 address
_reader.GoBack(); //unread the '[', so that we can start looking for ipv6
return(false);
}
else if (originalLength > 0 || numSlashes > 0 || !CharUtils.IsAlpha(curr))
{
// if it's not a character a-z or A-Z then assume we aren't matching scheme, but instead
// matching username and password.
_reader.GoBack();
return(ReadUserPass(0));
}
}
return(false);
}
