Added deprecated attribute API lint, minor incompatibility level

[epage]

Semver says deprecations are minor changes.

Reference: https://doc.rust-lang.org/reference/attributes/diagnostics.html#the-deprecated-attribute

Will require separate checks for:

• struct, enum
• union
• trait: new lint: trait_marked_deprecated #1092
• fn: new lint: function_marked_deprecated #1093
• const / static (GlobalValue): new lint: global_value_marked_deprecated #1101
• trait methods: new lint: trait_method_marked_deprecated #1098
• ImplOwner (struct/enum/union) methods: new lint: type_method_marked_deprecated #1130
• ImplOwner associated constants: new lint: type_associated_const_deprecated #1131
• trait associated types: new lint trait_associated_type_marked_deprecated #1099
• trait associated constants: new lint: trait_associated_const_marked_deprecated #1100
• enum variants: new lint: enum_variant_marked_deprecated #1179
• fields within enum variants: feat: added new lint enum_variant_field_deprecated #1096
• struct fields: new lint: struct_field_marked_deprecated  #1125
• declarative macros: new lint: macro_marked_deprecated #1102
• proc macros: new lint: proc_macro_marked_deprecated #1118

[pksunkara]

@obi1kenobi It hasn't been implemented for union. You might want to separate it out from struct & enum in the issue description.

[obi1kenobi]

Done, thanks!

[Frank-III]

struct fields should be marked done: #1125. the only two missing parts have PR opened.

[obi1kenobi]

Sweet! Thank you for all your work, that's a lot of boxes you've checked!

Might you be interested in doing a bit of schema extension next, to unlock even more lints elsewhere? If so, adding "field ordering" information to struct fields and enum tuple/struct variant fields can unlock ~10ish more new lints. This is because changing the order of struct fields (e.g. swapping which field is first and which is second) can be a major breaking change in some circumstances.

[Frank-III]

Sweet! Thank you for all your work, that's a lot of boxes you've checked!

Thank you for your guidance, the more lint I added the more I realized how important c-s-c would be in the rust ecosystem!

Might you be interested in doing a bit of schema extension next, to unlock even more lints elsewhere? If so, adding "field ordering" information to struct fields and enum tuple/struct variant fields can unlock ~10ish more new lints. This is because changing the order of struct fields (e.g. swapping which field is first and which is second) can be a major breaking change in some circumstances.

Definitely yes. Could you explain more about how it would be implemented? like adding a field to the StructField / Variant type ? and for the lints, we need to check if say a struct's field's order is changed, and report something like:

• "field a in struct order has changed"
• "field a in struct order has changed"

[obi1kenobi]

Here's an issue describing this in more detail and giving one example of a lint: obi1kenobi/trustfall-rustdoc-adapter#758

There are more lints that use order info in #5. I'll try to organize them into a separate issue with individual checkboxes similar to this one.

Off-topic: You are very good at this work! Consider applying for a Rust-themed Google Summer of Code of project, if you don't already have other plans this summer! There are some cargo-semver-checks projects on that list, and you can always pitch your own as well. Happy to chat more if that sounds interesting!

[Frank-III]

Off-topic: You are very good at this work! Consider applying for a Rust-themed Google Summer of Code of project, if you don't already have other plans this summer! There are some cargo-semver-checks projects on that list, and you can always pitch your own as well. Happy to chat more if that sounds interesting!

Wow, I did not know that! would definitely apply for it and see what I could build for the summer, c-s-c seems to be something I wanna work for!