Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Knex migration attempts to create already existing changelog and fails entire migration #2499
Since the upgrade to 0.14.4 we've started getting this error in CI:
It looks like 7ff766f actually wasn't as innocent as expected.
I think that syntax is not supported, but there has been requests to support something like that. Basically the problem is probably that
Probably it is necessary to change configuration to also support
Need to dig up old discussions and check out the code to be able to decide exactly how to support that.
Former did support schema only by accident. It has never been an official tested / documented feature. Schema reference should have been given currently with
Reason for not supporting
Anyways I'am interested in looking through all the knex APIs and see if it would be feasible to allow, test and document
@nunorafaelrocha no harm done, just mentioned about dangers of using implementation deatails as features.
This breaking/non-breaking/semver discussion has actually been discussed here at least couple of times before :) Knex is following semver and Knex is still < 1.0 so semver says that every change can be breaking change (of course it is debateable if knex version 1.0 should already have been released).
That aside line between non-breaking change and non breaking ones is pretty shady, since every small change that can be tested also changes functionality in some way and may break someone's code (even making simple bugfixes / adding new attribute somewhere can break someone's tests). So mostly I've been drawing that line where changing functionality that is not documented/tested public API is non-breaking change and changing documented functionality is breaking.
That being said of course if I would have known that people is using this implementation detail here, I would have marked it as breaking one even that it was undocumented functionality.
@igor-savin-ht nice I'm looking forward to check that PR