Commits on Source (35)
-
gjahn authored
-
gjahn authored
Hint in readme on how to change built target platform.
-
gjahn authored
-
gjahn authored
For now it's just ephemeral. Instead of disabling the timeout (-1), it's set to 1 ms for all requests in the state paths test. Otherwise race condistion may occur. Symptoms include incorrect state name, because it may have been read before it's fully written to the store.
-
gjahn authored
-
gjahn authored
This is being handled by Fiber, hence not tested.
-
gjahn authored
The version is displayed on the HTML-based landing page.
-
gjahn authored
Available levels for now: debug, error. While the former is the most verbose one and the latter the least verbose. It sets it to 'debug' if in development mode.
-
gjahn authored
* rework indentation of some configuration variables * rename some 'err' variables to '_' since they are not used
-
gjahn authored
-
gjahn authored
This is to improve/funnel potentially fatal errors
-
gjahn authored
-
gjahn authored
Using the Param "name" to create an Item and persist turns out to be affected bi Fiber's behaviour of "zero allocation", which means values may be re-used across requests/contexts. This caused wrong or chunked up Item names in the ephemeral state, which only surfaced during unit testing. For more details see https://docs.gofiber.io/#zero-allocation
-
gjahn authored
Initially, that was done to prevent flaky test runs, which turned out to be caused by Fiber's behaviour (se previous commit).
-
gjahn authored
-
gjahn authored
Somebody havent heard about SemVer yet. The Ctx.GetReqHeaders() return value slightly changed (string to []string), which now requires to first join the slice before checking whether it contains the given string
-
gjahn authored
Fore more information see https://go.dev/blog/maps#concurrency
-
gjahn authored
The location where the actual items should reside is in the store, while only references are being returned when attempting to access an item.
-
gjahn authored
-
gjahn authored
THis change set aims to simulate running test as if they would be executed in a pipeline
-
gjahn authored
* remove some spaces * return item.Name() instead of map key * re-order router log output
-
gjahn authored
-
gjahn authored
-
gjahn authored
-
gjahn authored
-
gjahn authored
-
gjahn authored
-
gjahn authored
-
-
gjahn authored
-
gjahn authored
* new slog in stdlib is utilized for that and replaces 'log' pkg * configuration pkg is used validate and now set/define/refer log level values * the 'error' level call is not fatal, thus os.Exit is called instead
-
gjahn authored
Although, password file is checked during config validation, sth may have changed the file content in the short period of time between validation and DB client initialization, hence exiting in case of an error
-
gjahn authored
-
gjahn authored
Showing
- .gitlab-ci.yml 11 additions, 0 deletions.gitlab-ci.yml
- Makefile 10 additions, 1 deletionMakefile
- README.md 135 additions, 0 deletionsREADME.md
- configuration/config.go 86 additions, 4 deletionsconfiguration/config.go
- go.mod 8 additions, 5 deletionsgo.mod
- go.sum 16 additions, 0 deletionsgo.sum
- main.go 35 additions, 3 deletionsmain.go
- routing/routes.go 223 additions, 6 deletionsrouting/routes.go
- routing/routes_test.go 135 additions, 11 deletionsrouting/routes_test.go
- routing/templates.go 11 additions, 0 deletionsrouting/templates.go
- state/ephemeral.go 86 additions, 0 deletionsstate/ephemeral.go
- state/ephemeral_test.go 37 additions, 0 deletionsstate/ephemeral_test.go
- state/item.go 30 additions, 0 deletionsstate/item.go
- state/persistent.go 119 additions, 0 deletionsstate/persistent.go
- state/store.go 12 additions, 0 deletionsstate/store.go
.gitlab-ci.yml
0 → 100644
README.md
0 → 100644
state/ephemeral.go
0 → 100644
state/ephemeral_test.go
0 → 100644
state/item.go
0 → 100644
state/persistent.go
0 → 100644
state/store.go
0 → 100644