Support isize and usize #72
Conversation
| @@ -212,6 +213,8 @@ impl Formatter { | |||
| /// leb64-encode `x` and write it to self.bytes | |||
| #[doc(hidden)] | |||
| pub fn leb64(&mut self, x: u64) { | |||
| // FIXME: Avoid 64-bit arithmetic on 32-bit systems. This should only be used for | |||
| // pointer-sized values. | |||
There was a problem hiding this comment.
could you create a low priority issue for this work?
There was a problem hiding this comment.
Hmm, timestamps are u64 and use leb64, so this isn't quite correct. Should we make timestamps usizes?
There was a problem hiding this comment.
leb64 is LEB128 with 64-bit arithmetic. The 10 byte buffer should be large enough for 64 bit values, no? Why make the timestamps usize which are 32-bit values on the target side?
There was a problem hiding this comment.
Yeah that works, but if leb64 would only be used on usize-sized values, we could avoid 64-bit arithmetic as well and just do it with 32-bit (assuming a 32-bit target). This is currently almost the case, except that the 64-bit timestamps also use leb64.
There was a problem hiding this comment.
I see what you mean. For alpha release I think we want 64-bit timestamps working.
No description provided.