1 unstable release
0.1.0 | Jun 3, 2021 |
---|
#315 in Internationalization (i18n)
26 downloads per month
Used in 2 crates
3KB
i18n-rs
A simple compile time i18n implementation in Rust.
This is a personal project. If you need a stable and powerful i18n library, you may need fluent.
If you think this crate is not easy to use, I found another similar crate: https://github.com/terry90/internationalization-rs
Use
In crates.io, the name of this package is simple-i18n
, because the name of i18n-rs
is occupied by an empty crate. shit...
Add simple-i18n = "0.1"
to Cargo.toml
Examples
Look i18n-example
cd examples/i18n-example
LOCALE_PATH=locale cargo run --package i18n-example --bin i18n-example
Docs
i18n-rs will load your locale (toml, json or yaml) into the code during compilation.
Then you can use lang!
(to switch the locale) and use i18n
to get the text.
LOCALE_PATH
The LOCALE_PATH
environment variable is used to find your locale file.
Please note that because the dependent library cannot get the current path where you actually run the build command during compilation, the safest method is actually to use the absolute path.
Usually we will store locale files in the locale
directory in the project root directory and set LOCALE_PATH
to locale
.
The current behavior is:
i18n-rs will find
${workspace_root}/$LOCALE_PATH
inOUT_DIR
or./
usingcargo metadata
.
In other words, if
LOCALE_PATH
is a relative path, it should be based on the workspace_root of the project, not the user's current path.
And it is best to run the build command in the project root directory.
Locale files
The locale file supports json
, toml
, yaml
.
You can use a single file or use a folder. In the case of a single file, the language code is the file name.
In the case of a folder, the folder name is language code, and the subfolder and file name will be used as the field name.
You can add .
in front of the file name to avoid becoming a field name.
The content will be flattened, and the key will be linked together with .
to become the field name.
Example:
{
"words": {
"greetings": {
"hi": "Hi!"
}
}
}
equal
{
"words.greetings.hi": "Hi!"
}
Strict and Loose
By default, strict checking will be used.
In loose mode, if you try to get a non-existent field or a non-existent locale, the field itself will be returned.
But strict mode will check your input in lang!
and i18n!
to make sure that you are using the existing locale and fields that exist in all locales.
If there is an error, it will be panic!
.
Don't worry, all of this is checked at compile time, so strict checking will hardly affect runtime performance, and there will be not panic at runtime.
note: Because it needs to be checked at compile time, string literals must be used in strict mode
Fortunately, We can freely switch between loose and strict mode.
like i18n!("xxx.x"; loose)
.
Benchmark
strict contrast/no strict
time: [29.048 ns 29.387 ns 29.736 ns]
change: [-15.897% -13.053% -10.253%] (p = 0.00 < 0.05)
Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high mild
strict contrast/strict time: [29.108 ns 29.431 ns 29.776 ns]
change: [-2.6412% -0.8426% +1.0984%] (p = 0.38 > 0.05)
No change in performance detected.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) high mild
2 (2.00%) high severe
change_lang time: [148.38 ns 159.76 ns 178.01 ns]
change: [+0.4039% +4.5240% +10.326%] (p = 0.05 > 0.05)
No change in performance detected.
Found 5 outliers among 100 measurements (5.00%)
3 (3.00%) high mild
2 (2.00%) high severe