Mario Ranftl

Hi @bar-ailo! Oh, it seems we forgot to document the path prefixes properly, see [templates router](https://github.com/allaboutapps/integresql/blob/master/internal/api/templates/routes.go) und [admin router](https://github.com/allaboutapps/integresql/blob/master/internal/api/admin/routes.go). Please prepend `/api/v1/templates` to all calls except deleting all templates (which...

Hi wouterhund, I see, our docs are misleading, we need additional clarification. `DELETE /templates/{hash}/tests/{test-database-id}` is *optional* and not going to destroy the test-database. Instead, it just *deletes* the **`dirty`** lock,...

Hi, see [/pkg/manager/manager.go `createNextTestDatabase`](https://github.com/allaboutapps/integresql/blob/master/pkg/manager/manager.go#L507). We currently just start by recreating the first encountered `dirty` database after the pool is exceeded. We currently cannot and do not want to know if...

After taking a closer look into the recreation mechanism, I can no longer confirm that this is a bug. The buffer mechanism is based on the `INTEGRESQL_TEST_INITIAL_POOL_SIZE` env var, meaning...

Thanks for your investigation, I understand your concern that the locking mechanism used here is suboptimal for selecting the next ready test database. Nevertheless, are you sure that you are...

> 1. Create a new test database > 2. Close all connections to the previous test database Yep, this could actually have an impact, as we try to destroy database...

Hi, this might be related to [Go modules @v1.13](https://maelvls.dev/go111module-everywhere/#go111module-with-go-113), anything >v1.14 should be fine, just tried the following inside a [golang:1.16.3](https://hub.docker.com/_/golang) container: ```bash $ go get github.com/allaboutapps/integresql/cmd/server go: downloading github.com/allaboutapps/integresql...

Hi @FedorchukAndrii . Thanks for opening this issue. Unfortunately we currently do not support Apples M1/M2 or the ARM architecture yet, but will be looking into this in the future...

Hm, restarted. Works, potentially the currently poor-mans caching implementation is insufficient with the current traffic load (designed ephemeral, however dyno is never restarted).

hm, I'm no longer 100% sure how an empty `local('')` got in there in the first place... Seems like it's there to support IE6-8, see this [stackoverflow issue](https://stackoverflow.com/questions/3698319/css-font-face-what-does-src-local-mean). Looks like...