feat(macros): Upgrade syn to v2 #1559
Conversation
| [dependencies] | ||
| bae = { version = "0.1", default-features = false } | ||
| syn = { version = "1", default-features = false, features = ["parsing", "proc-macro", "derive", "printing"] } | ||
| bae2 = { version = "1", default-features = false } |
There was a problem hiding this comment.
Since bae crate is no longer maintained and it is required that bae is upgrade to syn v2 too, so I made a fork of bae and published to crates.io.
There was a problem hiding this comment.
Is there any alternative for bae?
There was a problem hiding this comment.
@billy1624
I think we can adopt darling after TedDriggs/darling#226 got merged. Also it is not so difficult to parse attributes without bae or darling, thanks to parse_nested_meta API. In fact, major of sea_orm attributes are parsed manually without using bae.
There was a problem hiding this comment.
I was thinking we use Darling since it has nicer errors when you get a name wrong
| continue; | ||
| } | ||
| } else { | ||
| if !attr.path().is_ident("sea_orm") { |
There was a problem hiding this comment.
We can use this shorthand syntax.
| if let Expr::Lit(exprlit) = &nv.value { | ||
| if let Lit::Str(litstr) = &exprlit.lit { |
There was a problem hiding this comment.
The value of MetaNameValue is now Expr, so we need to check whether the value is Lit or not first.
|
Hey @siketyan, thanks for the hard work!! Perhaps we can rewrite all macros after TedDriggs/darling#226 merged and released? I will convert this into a draft PR for now. |
|
@billy1624 Yes, maybe. I saw your PoC. But I do not know we can or can not rewrite Btw I can work on this after the syn's PR got merged. |
|
+1 on using darling for attributes. they seem to have more helpful error messages for mistyped attribute names |
|
I totally agree to use darling instead. TedDriggs/darling#226 has been merged and we're ready to migrate. |
3 participants
PR Info
New Features
N/A
Bug Fixes
N/A
Breaking Changes
N/A
Changes