2016-02-09 15:32 GMT+01:00 Marko Tiikkaja <marko@joh.to>:
On 08/02/16 14:16, Pavel Stehule wrote:
2016-02-08 13:53 GMT+01:00 Marko Tiikkaja <marko@joh.to>:
Yeah, and that's exactly what I don't want, because that means that CREATE SCHEMA VARIABLE suddenly breaks existing code.
theoretically yes, but this conflict can be 100% detected - so no quiet bug is possible, and plpgsql_check can find this issue well. If you don't use schema variable, then your code will be correct. You have to explicitly create the variable, and if there will be any problem, then the problem will be only in PL functions in one schema. And you can identify it by statical analyse.
I'm sorry, but I think you've got your priorities completely backwards. You're saying that it's OK to add a footgun because blown-off pieces of feet can be found by using a third party static analyzer barely anyone uses. And at best, that footgun is only a very minor convenience (though I'd argue that omitting it actually hurts readability).
I don't block the integration plpgsql_check to upstream. I spent hundreds hours for it.
Can we look on this problem with different side? What I can do it for safe using proposed schema variables.
The possible ways:
1. requirement prefix like : or @. I don't prefer it because a) hard to find a agreement - Oracle fans like ":", MSSQL like @, other maybe $, b) with any currently unsupported syntax I have to fix SQL lexer, parser
2. requirement to use qualified name everywhere - it can works, but I don't prefer it, because sometimes can be unfunny to write long qualified identifiers. There are not aliases on schema in PLpgSQL. Possible solved by variable aliases. But it requires alias.
3. plpgsql GUC where schema variables are: a) disabled, b) enabled, c) only qualified names are allowed - it is similar to #variable_conflict option
I prefer @3 with "c" as default, but I can live with @2, and dislike @1 due mentioned reasons.