Option to preserve comment text when parsing templates

[jg-rp]

Comment text from {% comment %} and {% # .. %} tags is currently not included in a template's abstract syntax tree.

The translation message extractor under development in the Liquid Babel project needs access to comment text when building message catalogues. Therefore, we need the option to store comments in CommentNode objects for later static analysis.

Non-standard shorthand comment syntax gets stripped out by the lexer. For now, we'll say shorthand comments are not supported by the translation message extractor.

[jg-rp]

I've decided to call this a bug, which has been fixed in version 1.4.4.

The default behaviour is now to keep comment text in a template's AST. If anyone needs to save a bit of memory by excluding comment text, you can register liquid.builting.tags.comment_tag.CommentTag with a liquid.Environment, then load and render templates from that environment.

from liquid import Environment
from liquid.builting.tags.comment_tag import CommentTag

# This will override the default `CommentTextTag` with one that discards comment text.
env = Environment()
env.add_tag(CommentTag)

template = env.from_string("{% comment %}hello{% endcomment %}")