filters: escape arguments to pluralize at compile time
#174
Conversation
Kijewski
force-pushed
the
pr-compile-time-pluralize
branch
from
d13b5fc
to
5b9a623
Compare
Kijewski
force-pushed
the
pr-compile-time-pluralize
branch
from
5b9a623
to
d06480f
Compare
Kijewski
force-pushed
the
pr-compile-time-pluralize
branch
from
d06480f
to
5c34511
Compare
Kijewski
force-pushed
the
pr-compile-time-pluralize
branch
from
5c34511
to
2744364
Compare
rinja_derive/src/generator.rs
Outdated
| buf.write(format_args!("{CRATE}::filters::pluralize(")); | ||
| self._visit_arg(ctx, buf, count)?; | ||
| for value in [sg, pl] { | ||
| buf.write(", "); |
There was a problem hiding this comment.
| buf.write(", "); | |
| buf.write(','); |
rinja_derive/src/generator.rs
Outdated
| return None; | ||
| }; | ||
| match kind { | ||
| Some(IntKind::I8) => is_signed_singular(i8::from_str_radix, value, 1, -1), |
There was a problem hiding this comment.
I'm trying to think of a way to not have this huge match because it's very error prone to have to ensure that everything is working out correctly... Isn't it be possible somehow to have an impl with an associated type? Another (easier) way is to write a macro where we list types and then everything is generated correctly.
There was a problem hiding this comment.
I added a macro, but it's not much less ugly.
There was a problem hiding this comment.
Indeed but at least less error-prone, well done!
rinja_derive/src/generator.rs
Outdated
| Some(IntKind::I64) => is_signed_singular(i64::from_str_radix, value, 1, -1), | ||
| Some(IntKind::I128) => is_signed_singular(i128::from_str_radix, value, 1, -1), | ||
| Some(IntKind::Isize) => { | ||
| if cfg!(target_pointer_width = "16") { |
There was a problem hiding this comment.
Why not doing isize::from_str_radix?
There was a problem hiding this comment.
The target and host pointer widths may be different.
There was a problem hiding this comment.
Ah ok, cross-compilation was the issue.
|
Thanks for the improvement! |
|
Thanks! :) Should have thought about that already in the first PR. |
|
Well I definitely didn't think about that so that makes two of us. |
We can tell at compile time how to escape
""or"s". The former does not need to write anything, the latter does not need to be escaped.