netsurf: handle encoding change parser error

[krichprollsch]

When the HTML contains a META tag with a different encoding than the one used to parse the document, a c.DOM_HUBBUB_HUBBUB_ERR_ENCODINGCHANGE error is returned by netsurf.

In this case, we must restart the parsing with the new detected encoding.
The detected encoding is stored in the document and we can get it with documentGetInputEncoding().

Relates with #152 and #153

Slack discussion: https://lightpanda.slack.com/archives/C05TRU6RBM1/p1705070165019409

[krichprollsch]

A good solution would be to read the first 1kb of data from the reader, and try to extract the charset declaration in zig.
Then use it to parse the document instead of the HTTP one.

Then the parser must receive a custom reader which will contain the read buffer first and the following data after.