feat: Add a TryFrom<&str> implementation for enumerations #224
Conversation
| @@ -43,7 +68,7 @@ fn fmt_pascal_case(f: &mut std::fmt::Formatter<'_>, name: &str) -> std::fmt::Res | |||
| } | |||
|
|
|||
| macro_rules! lsp_enum { | |||
| (impl $typ: ty { $( $(#[$attr:meta])* pub const $name: ident : $enum_type: ty = $value: expr; )* }) => { | |||
| (impl $typ: ident { $( $(#[$attr:meta])* pub const $name: ident : $enum_type: ty = $value: expr; )* }) => { | |||
There was a problem hiding this comment.
not sure why, also the other : ty is still here
src/lib.rs
Outdated
| impl std::convert::TryFrom<&str> for $typ { | ||
| type Error = &'static str; | ||
| fn try_from(value: &str) -> Result<Self, Self::Error> { | ||
| match value { |
There was a problem hiding this comment.
I guess we can also write match ()
src/lib.rs
Outdated
| fn try_from(value: &str) -> Result<Self, Self::Error> { | ||
| match value { | ||
| $( | ||
| _ if { let (buf, len) = crate::fmt_pascal_case_const(stringify!($name)); &buf[..len] == value.as_bytes() } => Ok(Self::$name), |
There was a problem hiding this comment.
great idea. The only weird thing is that when I change the max length from 128 to 5 we will error at runtime (not at compile time!) if the value parameter is only known at runtime -- even though $name is known at compile time.
Assertions probably won't help here either (though they should be allowed in stable Rust soon)
So we might save allocations but we still run the code at runtime and don't generate the neat match-statement a human would write, which is "TypeParameter" => Ok(Self::TYPE_PARAMETER)
Anyway, that's not so important.
There was a problem hiding this comment.
So at this point it seems better to write
match pascal_case_to_screaming_snake_case(value) { "TEXT" => Ok(Self::Text), .. }
| _ if { | ||
| const X: (crate::PascalCaseBuf, usize) = crate::fmt_pascal_case_const(stringify!($name)); | ||
| let (buf, len) = X; | ||
| &buf[..len] == value.as_bytes() |
There was a problem hiding this comment.
pretty nice - this commit reduces the instruction count of the CompletionItemKind::try_from() from 4000 to just above 300 (in release mode, debug has around 10x more). Now let's hope that future rust will do this without the explicit const :)
The new next/prev symbol feature is typically used to jump around functions. We even provide a convenience wrapper with a default mapping. Some programming styles use mostly methods instead of functions. To make the default mapping more useful, extend it to jump to functions *or* methods, instead of just methods. Delete the convenience wrapper for methods, since that's now superseded. Swap the argument order of lsp-next-symbol and friends, so we can take a variable number of symbol kinds. Add a deserialization method until this feature becomes part of lsp_types, see gluon-lang/lsp-types#224
The new next/prev symbol feature is typically used to jump around functions. We even provide a convenience wrapper with a default mapping. Some programming styles use mostly methods instead of functions. To make the default mapping more useful, extend it to jump to functions *or* methods, instead of just methods. Delete the convenience wrapper for methods, since that's now superseded. Swap the argument order of lsp-next-symbol and friends, so we can take a variable number of symbol kinds. Add a deserialization method until this feature becomes part of lsp_types, see gluon-lang/lsp-types#224
The new next/prev symbol feature is typically used to jump around functions. We even provide a convenience wrapper with a default mapping. Some programming styles use mostly methods instead of functions. To make the default mapping more useful, extend it to jump to functions *or* methods, instead of just methods. Delete the convenience wrapper for methods, since that's now superseded. Swap the argument order of lsp-next-symbol and friends, so we can take a variable number of symbol kinds. Add a deserialization method until this feature becomes part of lsp_types, see gluon-lang/lsp-types#224
Closes #223
cc @krobelus