-
Notifications
You must be signed in to change notification settings - Fork 308
👊 👊 🍖 #4615
Comments
We have two sets of creds, one in |
VCR cassettes intermingle Braintree API calls with all of the other API calls, so if we want to rerecord all of the Braintree APIs then its work to do that without rerecording the other API calls. I mean, we rerecord whenever we CI master at Travis and that appears to be green so it should work, right? |
The problem I'm hitting is that the account represented by the creds currently in |
Why does it not work to rerecord existing tests against a new Braintree account? |
Oooh! Was there some special casing for |
|
Ah! Need the old |
Am I hitting the right account? |
I went to check production to see how it's set there. I guess I typed the password wrong because I got "Your account is now locked out. Please wait 30 minutes and try again." For one bad password attempt? 😳 |
False alarm! |
Oh! Need to re-record. 😊 |
💃 |
Okay! How many tests does that fix? 😁 |
Gah! Now hitting 'Gateway Rejected: duplicate', something I saw on #4613 and reported to Braintree support—don't expect that from sandbox requests. |
Workaround over there is purge sandbox (takes a few minutes). |
(And, helpfully, doesn't whack custom fields. 👍 ) |
(I've logged out of production. Better not go back in there tonight except hopefully to set the custom field!) |
Yeah, okay, back to these duplicates. :-/ I guess let's purge the sandbox and try again? |
/me kicks that off ... |
Okay, I'm back in. Rerunning ... |
Skunked.
|
|
Oh. 😁
|
|
Eight out of 14 fixture files have Braintree API calls in them. |
|
Hmm ... each fails locally when run individually ... both on |
Run together the two fail as at Travis.
|
Seems easy enough to record and add, but why didn't they get added in the first big re-record? |
Re-run of all of |
Repurposing this ticket to be my verbose log for shipping #4598 in the next ... three hours? |
Will see if #4616 runs green, then rebase the chain. |
Reviewing work to be done ... |
Done. Almost done organizing tickets ... |
Alright, so that's four more PRs needed against #4598:
Okay! 👊 👊 🍖 |
Debating whether to self-review-and-merge the four ready PRs on #4598 or start into new work and then review/merge all eight at once. Eight sounds like a lot, thinking let's get these four in now. |
I mean, Travis will take half an hour to run anyway. |
Which actually raises a good point: optimization is to merge the last in the chain that is ready to merge, to avoid the rebase-and-wait-for-Travis dance overhead. |
Re: dance overhead: https://www.youtube.com/watch?v=ovo6zwv6DX4. |
Alright, 4-chain updated, go Travis go! I know there are a couple things to fix on #4617, what about the others? |
Probably will need to bump #4609 back off in the interest of time. |
Errors, receipt, terms. Next chain! |
Sup with #4604 tho? |
Okay, fixed. I can do a simpler optional rationalization (no folding, just the two sections) on #4613. |
Alright, squeezing out last little bit of CI floofle on #4621 and then the new three-chain is ready for final review and merge. 👍 |
Landed #4598! |
Trying to get commits from #4613 into some kind of usable review pipeline. I'm in a mess with Braintree remote API creds during testing and VCR (re)recording.
The text was updated successfully, but these errors were encountered: