Re: ResourceOwner refactoring
От | Peter Eisentraut |
---|---|
Тема | Re: ResourceOwner refactoring |
Дата | |
Msg-id | 48b16f57-4204-4bd2-ab42-fb24144649c4@eisentraut.org обсуждение исходный текст |
Ответ на | Re: ResourceOwner refactoring (Heikki Linnakangas <hlinnaka@iki.fi>) |
Ответы |
Re: ResourceOwner refactoring
|
Список | pgsql-hackers |
It looks like this patch set needs a bit of surgery to adapt to the LLVM changes in 9dce22033d. The cfbot is reporting compiler warnings about this, and also some crashes, which might also be caused by this. I do like the updated APIs. (Maybe the repeated ".DebugPrint = NULL, /* default message is fine */" lines could be omitted?) I like that one can now easily change the elog(WARNING) in ResourceOwnerReleaseAll() to a PANIC or something to get automatic verification during testing. I wonder if we should make this the default if assertions are on? This would need some adjustments to src/test/modules/test_resowner because it would then fail.
В списке pgsql-hackers по дате отправления: