string.gsub() doesn't accept a pattern with zero byte in a middle

[Totktonada]

tarantool> string.gsub('\x00abc', '[%z\x01-\xff]', 'x')
---
- xxxx
- 4
...

-- Expected "xxxx".
tarantool> string.gsub('\x00abc', '[\x00-\xff]', 'x')
---
- error: malformed pattern (missing ']')
...

-- Expected "axxxd".
tarantool> string.gsub('ab\x00cd', 'b\x00c', 'xxx')
---
- "axxx\0cd"
- 1
...

Other functions that accept a pattern look affected by the problem too, at least string.find() is affected.

[Buristan]

Hi, Alexander!

As you can see from Lua 5.1 Reference Manual for string.find()

This function does not accept string values containing embedded zeros, except as arguments to the q option.

PUC RIO Lua accept embedded zeros in patterns since Lua 5.2.
So, we can implement this feature within full Lua 5.2 compatibility, if we want.

Also, see the same ticket in LuaJIT repo: LuaJIT/LuaJIT#759.

[Totktonada]

So, we can implement this feature within full Lua 5.2 compatibility, if we want.

We can implement it within partial Lua 5.2 compatibility as well, because the change is compatible.

As you can see from Lua 5.1 Reference Manual for string.find()

This function does not accept string values containing embedded zeros, except as arguments to the q option.

The quote is about string.format(). However you're right, the manual says the same for patterns:

A pattern cannot contain embedded zeros. Use %z instead.

So, technically speaking, this is not a bug. It is a feature request.