Oleg Kalnichevski wrote: > > If you want the parser to be truly generic you cannot expect the version > part to always be 4 char in length as it can also be something like > 20.121 or 234.5
Yes, I know. That branch will only be executed when someone is looking for a protocol version at the end of the line, which is not used anywhere in our code at the moment. The functionality is the result of a generalization I made, and I'm still considering to remove it rather than to add reverse scanning code. The realistic cases I see are HTTP/1.1, HTTP/1.0, SIP/2.0 If there is a real need for multi-digit version numbers, a derived parser can be used as a workaround. Good catch! cheers, Roland --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
