I'm unclear — how do we know a WHERE is unparseable? Have we made a too-clever fake? If we're trying to test the behaviour of queries, it would be better to use an actual clickhouse instance.
Yes — too-clever fake. Specifically:
| [Security] | |
| EAP-Method=PEAP | |
| EAP-Identity=anonymous | |
| EAP-PEAP-Phase2-Method=MSCHAPV2 | |
| EAP-PEAP-Phase2-Identity=USERNAME_GOES_HERE | |
| EAP-PEAP-Phase2-Password=PASSWORD_GOES_HERE |
| [GLOBAL] | |
| pants_version = "2.15.0" | |
| backend_packages = [ | |
| "pants.backend.python" | |
| ] | |
| [source] | |
| root_patterns = ["/"] | |
| [python] |
| def sip(stream): | |
| """ | |
| NAYA will handle our json just fine so long as it starts with an | |
| open bracket, so it thinks it has a list. | |
| """ | |
| def f(): | |
| # Yield a dummy `[` | |
| yield TOKEN_TYPE.OPERATOR, "[" | |
| # Followed by the rest of the tokens in the stream | |
| for t in tokenize(stream): |
| resource "aws_cloudwatch_log_group" "debug" { | |
| name = "API-Gateway-Execution-Logs_${aws_api_gateway_rest_api.stats.id}/${var.stage_name}" | |
| retention_in_days = 1 | |
| } | |
| resource "aws_iam_role" "apig_cw" { | |
| name = "stats-apig-cloudwatch-logger" | |
| assume_role_policy = <<-EOF | |
| { | |
| "Version": "2012-10-17", |
| import { Station, bill } from "../lib"; | |
| describe("A single journey in zone A", () => { | |
| const taps: Array<Station> = [Station.Asterisk, Station.Aldgate]; | |
| it("should charge £2.50 for the journey", () => { | |
| const [charge] = bill(taps); | |
| expect(charge.amount).toEqual(250); | |
| }); | |
| }); |
Although the pace of change at Made has slowed over the last 12 months, we are still testing new techniques and re-examining best practices. It's useful to periodically reflect on what good practice looks like and make sure we're spreading that knowledge to our dev teams. This is a quick guide to what the Ops team need from developers in 2018. If your systems don't comply with these guidelines, you should chat to a friendly ops person so we can help you prioritise and fix issues.
Ops can't operate your system if they don't know how to work it. Every system should have an operators.md file in the root of the github repository that describes:
I hereby claim:
To claim this, I am signing this object:
| def doSomethingThatMightFail(): | |
| attempts = 0 | |
| success = False | |
| retry = True | |
| while retry and not success: | |
| success = call_the_service() | |
| if attempts ++ > MAX_ATTEMPTS: |