Add new custom type Arg, update API to main! : List Arg => Result {} [Exit I32 Str]_
#293
Conversation
| // TODO confirm this is ok... feels dangerous | ||
| let os_str = | ||
| std::ffi::OsString::from_encoded_bytes_unchecked(c_str.to_bytes().to_owned()); |
There was a problem hiding this comment.
Please read the docs and give me an opinion here 🙏
There was a problem hiding this comment.
The byte encoding is an unspecified, platform-specific, self-synchronizing superset of UTF-8. By being a self-synchronizing superset of UTF-8, this encoding is also a superset of 7-bit ASCII.
So I think this is incorrect, but it also seems unnecessary. This seems like we could just pass the &[u8] to ArgToAndFromHost::from(arg) and handle it there. For Unix it's handled, and for Windows we'll need to do something special anyway.
| @@ -3,7 +3,7 @@ app [main!] { pf: platform "../platform/main.roc" } | |||
| import pf.Stdout | |||
| import pf.Cmd | |||
|
|
|||
| main! = \{} -> | |||
| main! = \_ -> | |||
There was a problem hiding this comment.
If this was personal code, then ignoring the args entirely would probably be expected, but we should default to something more instructive in these examples, meaning I vote for _args instead of just _ for all these examples:
- It tells readers what is being ignored, so they know that
main!gets a list of arguments even if they only read the one example - It implies that naming ignored variables is good practice (which it generally is) to give context to readers
| Unix -> Unix inner.unix | ||
| Windows -> Windows inner.windows | ||
|
|
||
| from_os_raw : InternalArg.ArgToAndFromHost -> Arg |
There was a problem hiding this comment.
| from_os_raw : InternalArg.ArgToAndFromHost -> Arg | |
| ## Wrap a raw, OS-aware numeric list into an [Arg]. | |
| from_os_raw : InternalArg.ArgToAndFromHost -> Arg |
| ## Unwrap the raw bytes for decoding, typically this is | ||
| ## consumed by a package and not an end user | ||
| to_os_raw : Arg -> [Unix (List U8), Windows (List U16)] |
There was a problem hiding this comment.
A U16 isn't considered a byte, so we can't call the data type "bytes".
| ## Unwrap the raw bytes for decoding, typically this is | |
| ## consumed by a package and not an end user | |
| to_os_raw : Arg -> [Unix (List U8), Windows (List U16)] | |
| ## Unwrap an [Arg] into a raw, OS-aware numeric list. | |
| ## | |
| ## This is a good way to pass [Arg]s to Roc packages. | |
| to_os_raw : Arg -> [Unix (List U8), Windows (List U16)] |
| # os_str's are not necessarily valid utf-8 or utf-16 | ||
| # so we store as raw bytes internally to avoid | ||
| # common mistakes | ||
| Arg := InternalArg.ArgToAndFromHost |
There was a problem hiding this comment.
| # os_str's are not necessarily valid utf-8 or utf-16 | |
| # so we store as raw bytes internally to avoid | |
| # common mistakes | |
| Arg := InternalArg.ArgToAndFromHost | |
| ## An OS-aware representation of a command-line argument. | |
| ## | |
| ## Though we tend to think of args as Unicode strings, most operating systems | |
| ## represent command-line arguments as lists of bytes that aren't necessarily | |
| ## UTF-8 encoded. Windows doesn't even use bytes, but U16s. | |
| ## | |
| ## Most of the time, you will pass these to packages and they will handle the | |
| ## encoding for you, but for quick-and-dirty code you can use [display] to | |
| ## convert these to [Str] in a lossy way. | |
| Arg := InternalArg.ArgToAndFromHost |
| from_os_raw : InternalArg.ArgToAndFromHost -> Arg | ||
| from_os_raw = \raw -> @Arg raw | ||
|
|
||
| ## Convert an Arg to a `Str` for display purposes |
There was a problem hiding this comment.
| ## Convert an Arg to a `Str` for display purposes | |
| ## Convert an Arg to a `Str` for display purposes. |
| Unix -> Unix inner.unix | ||
| Windows -> Windows inner.windows | ||
|
|
||
| from_os_raw : InternalArg.ArgToAndFromHost -> Arg |
There was a problem hiding this comment.
Also, this should probably take [Unix (List U8), Windows (List U16)] to not expose internals.
| // TODO confirm this is ok... feels dangerous | ||
| let os_str = | ||
| std::ffi::OsString::from_encoded_bytes_unchecked(c_str.to_bytes().to_owned()); |
There was a problem hiding this comment.
The byte encoding is an unspecified, platform-specific, self-synchronizing superset of UTF-8. By being a self-synchronizing superset of UTF-8, this encoding is also a superset of 7-bit ASCII.
So I think this is incorrect, but it also seems unnecessary. This seems like we could just pass the &[u8] to ArgToAndFromHost::from(arg) and handle it there. For Unix it's handled, and for Windows we'll need to do something special anyway.
No description provided.