*Multline* comments getting flagged with fct_hard_coded_references

[guyr-ziprecruiter]

While fixed for ordinary comments:

dbt-project-evaluator/macros/find_all_hard_coded_references.sql

Lines 17 to 20 in 9f9c1db

	{# we remove the comments that start with -- , or other characters configured #}
	{%- set re = modules.re -%}
	{%- set comment_chars_match = "(" ~ var('comment_chars') | join("|") ~ ").*" -%}
	{%- set model_raw_sql_no_comments = re.sub(comment_chars_match, '', model_raw_sql) -%}

Mentioning table names inside multline comments

/*
my_database.my_table
*/

will result in a failure.

For now I switched from multiline SQL comments to jinja2 comments (which are multiline by default)

Switching the jinja2 multiline comments also does not cut it and gets flagged.

{#
my_database.my_table
#}

Fixing this could be nice 🙇 . Thanks!

[b-per]

I think this is where we reach the limits of what we can do with some already complex (but not insanely complex) regex.

I'd say that the best here is either to add the line comment in your multiline comment

/*
this is some comment
-- my_database.my_table
and this is some comment as well
*/

or use Jinja comments like you did

or convert your multi-line comments to multiple single-line ones

---

We could technically fix your specific use case, but I don't really want to then have to maintain it for all the combinations of multiline comments for every supported data warehouse