added 412 to error codes that are checked

[Mia-Cross]

The validation function for the cassettes will now check for 412 errors (Precondition failed) in addition to 409 and 500. This will be useful when dealing with private networks among other things.

[codecov-commenter]

Codecov Report

Merging #1306 (359918a) into master (7d0430b) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1306   +/-   ##
=======================================
  Coverage   70.22%   70.22%           
=======================================
  Files         107      107           
  Lines       17514    17514           
=======================================
  Hits        12299    12299           
  Misses       4150     4150           
  Partials     1065     1065           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7d0430b...359918a. Read the comment docs.

[remyleone]

We got some 412 in the cassettes, I've checked with a grep. It is weird that the CI does not catch anything. Can you ensure that the test is working properly and that HTTP codes are indeed checked?

[Mia-Cross]

We found a bug in the cassettes validation function that was preventing quite a few errors to show up, here's the detail of the errors that we can catch now :

[Mia-Cross]

As of today, we still have these errors to fix :

• code 500 : rdb-instance-private-network-dhcp.cassette
  --> at the beginning of the test, in response to POST request on the private networks route. I don't understand why since the GET requests made after this on the private network don't seem to fail.

• code 412 : rdb-instance-private-network.cassette & rdb-instance-private-network-dhcp.cassette
  ---> at the very end of the tests, in response to DELETE request on the second private_network, and this time adding the extra step of configuration with just the private network to wait for the cluster to be destroyed doesn't seem to do the trick. test(rdb_instance): detach private network in final steps #1335

• code 409 : rdb-database-manual-delete.cassette
  --> in response to DELETE request on the user. It seems strange that, among all the resources to be deleted, the first DELETE request that is performed is this one since "depends_on" is specified. But since 409 errors are dealt with in the resourceScalewayRdbUserDelete function using retries and that the user ends up deleted in the end, maybe we could find a way to make the validate_cassettes function more flexible for this type of case ? fix(rdb_privilege): rdb_privilege not being deleted #1334