Let's stop this nightmare. The "solution" that uses a font hack that overrides the sematnics of Latin letters will never work as it should. Th eseparation of code points is necessary, even if this is just to show an URL containing Sinhalese letters in the domain name part (and without alternig the semantic of the dot, slash and colon separators). It will be inacceptable to have the "http://"; prefix isolated with a separate font just to be read correctly. from the rest of the URL. Inacceptable also beauecause it will alter the internals of international stadnards that are widely used. Inacceptable because Sinhalese domain names will remain separated from the proposed "romanizations".
That user really has a complete misunderstanding of the standard and severly lacks basic knowledge of the concepts. He shuld first read the definitions to see that what is in the standard is definitely not what he suposes by just looking at a simple basic chart (which is mostly informative and has very littel use for technical implementations). Reading the standard up to Chapter 3 (crequirements and convrormance) is absolutely necesarry for him. He won't make any progress to understand his own problems before reading it and criticizing constantly what he has never read ffor not understanding it... He should also read the introduction of the OpenType specifications which also use their own definitions (wsomething he is mixing as well). He must absolutely first understnad the character model and the separation between what is Unicode, what is a "abstract character", a "glyph", an encoding form, and the binary serialisation of an encoding into a stream of bytes, plus other concepts used by common protocols and languages such as transport syntaxes and alternate representations using things like character entities (in SGML, XML, HTML), or numerical escapes (e.g. in C/C++, PHP, Java, Ruby...) or string expressions using builtin/Standard functions in Basic...)
