With KCIDB SQL database moving to the Dashboard, it is much easier to start expanding on it. Let's add series as a unique identifier for the same test across checkouts. It should be easier and more clear to the user how to identify the same test in the history, so we can have an easier way to look for regressions:
By the same build/test we mean when the following parameters are same:
build: config + compiler + arch
test: hw-platform + build.config + build.compiler + arch + path
The series could be filled in at ingestion time once we complete the new ingester. The ingester can generate a 10-char word out of parameters as the series and store in in the DB together with the rest of the data.
This will be quite useful for searching back into history. Only the series will be necessary as parameter.
With KCIDB SQL database moving to the Dashboard, it is much easier to start expanding on it. Let's add series as a unique identifier for the same test across checkouts. It should be easier and more clear to the user how to identify the same test in the history, so we can have an easier way to look for regressions:
By the same build/test we mean when the following parameters are same:
build: config + compiler + arch
test: hw-platform + build.config + build.compiler + arch + path
The series could be filled in at ingestion time once we complete the new ingester. The ingester can generate a 10-char word out of parameters as the series and store in in the DB together with the rest of the data.
This will be quite useful for searching back into history. Only the series will be necessary as parameter.