Vladimir Sitnikov

>seems random function generate multiple value when access more than once! That seems about right for a random function, doesn't it?

I would say `function.cache.per.iteration` is a wrong approach to handle things. At the same time I don't understand the problem you have.

Could you share a reproducer script, expected and the actual output?

Based on my current understanding, "auto detect" is likely not an option as it might create subtle parsing issues. For instance, with "positional" (JDBC) style, one has to duplicate `?`...

> If there is any named parameter like :xxx then it should be PlaceholderStyle.NAMED, otherwise it's positional. The query might contain **no** placeholders at all in which case you won't...

Please, this is getting off-topic. The key question for this issue is whether spring-framework is willing to use non-standard extensions to JDBC provided by the drivers

Consider the following: * Even though Spring might have its own "SQL syntax", it is doomed to send repeated variables several times. For instance, if the query has something like...

@snicoll , I understand `NamedParameterJdbcTemplate` is a Spring-specific class. However, I do not fully understand reasons for declining a proposal of improving `NamedParameterJdbcTemplate` by leveraging driver features. For instance, Java...

Ok. Let me try that. However, I think it won't be added to KotlinPoet judging by https://github.com/square/kotlinpoet/pull/496#issuecomment-441340885

As you can see, https://github.com/square/kotlinpoet/issues/877 was closed, so WDYT regarding extended blocks in KotlinPoetDSL?