Question

Suppose an URL is encoded in a multi-byte character set where one of the characters in the multi-byte sequence could be between 0 and 127, i.e. an otherwise valid 7-bit ASCII character.

Example: The Japanese Shift_JIS character set, where the character カ would
be escaped as %83%4a. Now %4a is also the ASCII character J, so I could instead write %83J.

Would that be OK by the whatever standard(s) apply?

I’m not asking because I want to send URLs like this (although the latter saves a couple bytes), but whether I should accept those on the server side, i.e. whether it is standards-compliant and also, whether I can expect other servers to handle this in the same way.

Answer 1

I’m basing my answer on RFC 2396, as that is what’s being used by HTTP 1.1.

According to Section 2.1, there are 2 separate steps, the latter being optional:

1. URI character sequence->octet sequence
2. octet sequence->original character sequence

So the answer is: Yes, it’s OK.