Every `Build()` method now requires a `context.Context`. This is so the rolodex knows where to resolve from when locating relative links. Without knowing where we are, there is no way to resolve anything. This new mechanism allows the model to recurse across as many files as required to locate references, without loosing track of where we are in the process.
Signed-off-by: quobix <dave@quobix.com>
The work is now almost fully reflective, and *drumroll* the original order of the data is maintained when re-rendering. new elements pop in at the top of the object, but everything else will remain in the sequence it was originally added.
When building a document, everything that IS NOT a schema is auto-resolved in the model, this is very convenient because it creates a nice tree to explore and there is no need to perform lookups to when using `$ref` instead of inline definitions.
The issue is however being able to determine if the node was originally a reference or not, that data was lost, including the name of the reference used. This use case surfaced in issue #25, where the need to know what is and what is not referenced has different requirements for different applications.
This update now tracks that data and makes it available in `NodeReference` and `ValueReference` for every property.
Signed-off-by: Dave Shanley <dave@quobix.com>
The more in-depth we use extensions, the more likely is is that we need custom extensions. New `UnpackExtensions` function in the high package makes this easy. Low level models have been updated to support feature fully and docs added in README and examples as well.
Signed-off-by: Dave Shanley <dave@quobix.com>
Designs for handling multiple versions of objects have been set, seems clean and scalable. Generic functions for handling maps has been added also, which will cut down time moving forward.
Which has led to a new wider hashing capability for the low level API. hashing makes it very easy to determine changes quickly, without having to run comparisons to discover changes, could really speed things up moving forward.
This simple method gives the low API a super powerful and simple way to mutate the value of any node, which is then reflected in the root node, can than be serialized again and, voila! now our spec is editable.
working with k8s, stripe, petstore and locals. Speed is as good as I can make it at this point, not without further performance tunupes for memory consumption. less copying everywhere.
Same functions will power every other model. Found some opprotunities to add control over circular references, when and how to fire errors and what not. All in all, some robust code additions will improve the library significantly.
Using some designs unearthed from building the higher level model, I have brough that design down to the lower level to speed things up. It only took 8 years, but finally, I think I have mastered mult-threading. No more deadlocks, and no more need for waitgroups for everything.
Time to revisit each model and build individual tests to ensure all error handling is in place across the model.
Signed-off-by: Dave Shanley <dave@quobix.com>
