Shopify Liquid allows for HTML encoded quotes where this package does not

[allenap]

For example:

from liquid import parse

# ' is the HTML entity for a single quote.
# This arises naturally in HTML files where Liquid tags appear inside
# HTML attribute values that have been entity-encoded by a renderer
# (e.g. React Email).
template = """{% if x == 'foo' %}bar{% endif %}"""

parse(template)  # --> exception

vs.

require "liquid"
Liquid::Template.parse("{% if x == 'foo' %}bar{% endif %}")

[jg-rp]

Hi @allenap,

Python Liquid aims for maximum compatibility with Shopify/liquid in strict mode. If we render your example in strict mode, Shopify/liquid raises a syntax error too.

require "liquid"

puts Liquid::Template.parse("{% if x == 'foo' %}bar{% endif %}", error_mode: :strict)

# Liquid syntax error: Unexpected character & in "x == 'foo'" (Liquid::SyntaxError)

Unfortunately the documentation explaining this got lost when moving to MkDocs with Python Liquid v2 (my fault, I'll update the docs shortly). From archived docs:

Python Liquid might not handle syntax or type errors in the same way as the reference implementation. We might fail earlier or later, and will almost certainly produce a different error message.

Python Liquid does not have a "lax" parser, like Ruby Liquid. Upon finding an error, Python Liquid's lax mode simply discards the current block and continues to parse/render the next block, if one is available. Also, Python Liquid will never inject error messages into an output document. Although this can be achieved by extending BoundTemplate and overriding render_with_context().

That being said, I'm open to the idea of adding a lax parser, at least for common cases like your example.

(Edit: Digging into Shopify/liquid a bit, lax mode parses 'foo' as x27, and discards everything after. So unless you have a variable called x27 in scope, the condition will always evaluate to false.)