Autodetect issues

[platima]

Hey I've found a few issues with autodetect, and wondering if I am doing something wrong or if there are limitations I don't understand? Primarily, many many things autodetect as 'C'.

Eg
WITH employee_ranking AS ( SELECT employee_id, last_name, RANK() OVER (ORDER BY salary DESC) as ranking FROM employee ) SELECT employee_id, last_name, FROM employee_ranking WHERE ranking <= 5 ORDER BY ranking

I'd expect to very clearly be SQL, and then
<meta property="og:description" content="One of the best ways to learn advanced SQL is by reading examples of queries. Check 25 advanced queries that will help you build your SQL skills." /> <meta property="og:type" content="article" /> <meta property="og:url" content="https://learnsql.com/blog/25-advanced-sql-query-examples/" /><meta property="article:published_time" content="2023-02-28T18:00:00&#43;02:00"/> <meta property="article:modified_time" content="2023-02-28T18:00:00&#43;02:00"/><meta property="og:site_name" content="LearnSQL.com" /> <meta property="article:publisher" content="https://www.facebook.com/groups/welearnsql/" /> <meta property="article:section" content="blog" /> <meta property="article:tag" content="sql" /> <meta property="article:tag" content="learn sql" /> <meta property="article:tag" content="online practice" /> <meta property="article:tag" content="advanced sql" />
is quite clearly HTML.

Both auto-detect as 'C' though. Tested with a few releases, same result. Putting at the start obviously works, but is not always how it'll be used.

Any ideas?

[ctapmex]

Hi.
You're using colorer.exe ? which version or source build (commit)?
the type definition is influenced by both the file name and the first lines - which name is used?

[platima]

Hey latest build, linux x64, and reading from stdin.

[ctapmex]

Auto-detection works on the file name and the first lines of the file. To do this, the prototype description must specify the appropriate regular expressions for them. If you look at proto.hrc, then there are no regular expressions for the first lines for the "database" group.
and since we use stdin, there is no file name for analysis either. For this reason, auto-detection takes the first file type in the list - others cannot defeat it.
Can you suggest regular expressions that would be suitable for the first lines of an arbitrary sql file?

[platima]

Ah roger that. I'll come up with some regex for a few and issue a PR

[platima]

colorer/Colorer-schemes#146

[ctapmex]

I'll roll out all the changes for now. Even the changes that remained broke the tests. In the meantime, there is no time to figure it out.
If you are interested to see what kind of tests, then https://github.com/colorer/Colorer-schemes/blob/master/README.md#develop

[platima]

Ah excellent, thanks!

[platima]

(FYI I did not update changelog sorry, might want to!)