Use Diplomat &[&str], cmp::Ordering, unvalidated UTF Writeable
#4786
Conversation
robertbastian
changed the title
&[&str]&[&str], cmp::Ordering
| ffi.Pointer<_SliceUtf8> allocIn(ffi.Allocator alloc) { | ||
| final slice = alloc<_SliceUtf8>(length); | ||
| for (var i = 0; i < length; i++) { | ||
| final codeUnits = Utf8Encoder().convert(_strings[i]); |
There was a problem hiding this comment.
Maybe extract the Utf8Encoder() to a field? It is even a const constructor, so writing const Utf8Encoder encoder could increase performance.
There was a problem hiding this comment.
Is that actually necessary? I'd hope the compiler doesn't need to create an empty object to call the method.
There was a problem hiding this comment.
I don't know if the compiler optimizes that, better to extract it and be sure IMO.
There was a problem hiding this comment.
prefer using const.
Const constructors may be used to get a constant, but can also be used to instantiate a new object. Using the const keyword ensures it is a const. (And will in AOT ensure the object is dropped if the method doesn't actually use the object.)
There was a problem hiding this comment.
A const field or a const variable?
robertbastian
changed the title
&[&str], cmp::Ordering&[&str], cmp::Ordering, unvalidated UTF Writeable
utils/writeable/src/utf.rs
Outdated
| out.push_str(valid); | ||
| out.push_str( | ||
| char::REPLACEMENT_CHARACTER | ||
| .encode_utf8(&mut [0; char::REPLACEMENT_CHARACTER.len_utf8()]), | ||
| ); |
There was a problem hiding this comment.
Suggestion: why not just use <String as core::fmt::Write>::write_char and write_str here?
There was a problem hiding this comment.
Because they're fallible
There was a problem hiding this comment.
Suggestion: use String::push(char) for the second one.
There was a problem hiding this comment.
Hmm, that has an extra branch which might not be optimized out. Probably cheapest to just use push_str
There was a problem hiding this comment.
Confirmed in godbolt that it does in fact get optimized out.
Co-authored-by: Shane F. Carr <shane@unicode.org>
| ); | ||
|
|
||
| out.push_str(valid); | ||
| out.push('�'); |
There was a problem hiding this comment.
Nit, optional: I would prefer explicitly writing out the code point so that I know this is a literal replacement character, U+FFFD, and not some other weird Unicode question mark symbol.
| out.push('�'); | |
| out.push('\u{FFFD}'); |
I say this aware of the following section in the style guide.
There was a problem hiding this comment.
No, I recognise the question mark as the replacement character, I don't recognise the escape sequence.
No description provided.