feat: require string tokens used as comparison values to be delimited

[jimlambrt]

Originally, you could could write a query like: name=alice

This PR requires that same query to be written as: name="alice"

This change will allow us to introduce new keywords and/or reserved words as needed in the future without introducing breaking changes. It also disambiguates symbols from strings.

Also, this PR sets us up to support additional string delimiters in a subsequent PR. For instance, we can support ` (backtick) and ' (singe quote) as string delimiters which would really improve the UX for writing queries.

This PR also includes:

• feat: add support to escape double quotes
• refact: split lexStringState() into two distinct state functions
• refact (lexSymbolState): eliminate the bytes.Buffer

[jimlambrt]

I still need to update the docs and add a CHANGELOG entry. Besides those chores, this PR is ready for review.

[jefferai]

Also, this PR sets us up to support additional string delimiters in a subsequent PR. For instance, we can support ` (backtick) and ' (singe quote) as string delimiters which would really improve the UX for writing queries.

Do we want this? Would it be better to standardize on a single delimiter so that you can include others without escaping? e.g. if we standardize on single quote, we don't need to escape embedded double quotes. Or if we standardize on double quotes we don't need to escape apostrophes.

[jimlambrt]

The idea is to make it easier for the user by allowing them to choose the most convenient delimiter for the query they're writing.

Examples using double-quotes, single-quotes and backticks:

description=`Alice's Restaurant`  
description="Alice's Restaurant" 
description='This restaurant is "owned" by Alice'

[talanknight]

Looks good. Thanks!

[johanbrandhorst]

Could you add some quoted strings to the fuzz tests?