=== on 16 Jun 2016, 17:15:29 Ann Campbell wrote:
https://www.researchgate.net/publication/222432151_Semantic_errors_in_SQL_queries_A_quite_complete_list

=== on 10 Aug 2017, 15:27:47 Pierre-Yves Nicolas wrote:
I don't agree with:

____the use of a wildcard without the use of LIKE must also be an error.____

Such cases can be intentional and we should not raise a "bug" for them.

=== on 10 Aug 2017, 15:44:27 Ann Campbell wrote:
\[~pierre-yves.nicolas] are you saying that if I search for "foo*" without ``++LIKE++`` then I'm expecting a literal asterisk?

=== on 10 Aug 2017, 16:05:19 Pierre-Yves Nicolas wrote:
\[~ann.campbell.2] "%" is a wildcard, not "*".

This RSPEC has 2 cases:

* ``++LIKE 'foo'++`` has definitely a problem because 'foo' has no wildcard. Whether it's a bug or a code smell is debatable.
* ``++= 'foo%'++`` may be a valid condition, "%" is not a really strange character. Should we raise a bug on ``++= '100%'++``? I think that we may raise a code smell but not a bug.

=== on 10 Aug 2017, 16:27:19 Ann Campbell wrote:
Sounds like we need to split the rule [~pierre-yves.nicolas]. You want to start the Rule dogfood thread?

=== on 2 Mar 2018, 15:35:09 Pierre-Yves Nicolas wrote:
\[~ann.campbell.2] This rule seems to find a lot of issues on T-SQL projects. That makes the classification as "major bug" hard to justify. I created a T-SQL subtask to make it "minor". Another possibility is to change the rule to a code smell for all languages. Any opinion?

=== on 2 Mar 2018, 17:25:07 Ann Campbell wrote:
https://stackoverflow.com/questions/543580/equals-vs-like