-
Notifications
You must be signed in to change notification settings - Fork 24.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Deprecate camelCasing #8988
Comments
+1 |
1 similar comment
+1 |
+1, what about deprecate in 1.5 and remove in 2.0? |
+1 to deprecate in 1.5 but remove in 2.1 or 3.0. |
So everyone seems to agree on this issue, the question is more about when the deprecation starts and when we remove. We can only deprecate as of 2.0 because of #8963 and then remove in 2.x. Even though it is a minor release, I would agree with David we should do it sooner rather than later, especially if we have it deprecated from the first 2.0 release? |
I don't think #8963 should stop us form deprecating (not having this warning didn't stop us before). Also, I don't think many ppl are using camelCasing or are even aware of it. Also it doesn't feel right removing something like this in between minor versions. +1 on deprecation for 1.5 and remove on 2.0 |
+1 from me as well to deprecate in 1.5 and remove in 2.0. |
This has been hanging around for way too long. Let's just bite the bullet and deprecate camel case in 5.0 (with deprecation logging) and remove it in 6.0 |
+1 . I think we should evaluate the BWC implications for stored data (percolator? mapping?) but I hope we can find ways to convert things …
|
In 5.0 mappings serialize themselves using lowercase (we do not just copy the user input) and the percolator now stores the binary representation of a query, so I think we are fine from this perspective. |
I meant things already stored in indices. Good to know we convert incoming input.
|
The current api allows for choosing which "case" response json keys are written in. This has the options of camelCase or underscore. camelCase is going to be deprecated from the query apis. However, with the case api, it is not necessary to deprecate, as users who were using it in 2.x can transition completely on 2.x before upgrading by simply using the underscore option. This change removes the 'case' option from rest apis. see elastic#8988
@rjernst can this issue be closed now? |
Looking up settings currently falls back to adjusting the setting key to camelCase, and then looking at the parsed settings keys again. This adds deprecation logging to that case. While in 5.0 settings validation will handle most of these cases, there stills exists some code (eg analysis providers) which lookup settings directly on the Settings object. see elastic#8988
@clintongormley As I was preparing the removal of all camel case from master, I found some other cases. One additional deprecation is in #17875, however, there are other places that I found are not using ParseField, and have their own custom parsing logic which includes many variations of settings names (including adding |
Now that the current uses of magical camelCase support have been deprecated, we can remove these in master (sans remaining issues like BulkRequest). This change removes camel case support from ParseField, query types, analysis, and settings lookup. see elastic#8988
@clintongormley Not sure I understand your comment about analyzer/etc. I added deprecation for the automatic addition of camel case analysis names in #17800. |
Elasticsearch allows all parameters to be passed with underscores or as camelCase (except where somebody has forgotten to handle the camelCase variant, and isn't using parseField #8964).
Also, the keys in responses can be rendered in camel case by passing
?case=camelCase
, except that rendering seems inconsistent, eg:Does anybody actually use the camelCase option? Almost all of the documentation uses the underscore version of parameters, and having a single set of accepted parameters (and responses) seems less likely to cause confusion than the current situation.
I'd recommend deprecating camelcase in 2.0 and removing it in 3.0, but labelling this ticket for discussion.
The text was updated successfully, but these errors were encountered: