Idiomatic way to convert non-UTF-8 vector slice to PathBuf

[Yogi39]

Please! I am sure this must be trivial, but I have spent too long trying to find the syntax to achieve this.

Thanks, Yogi39

[alice]

This is OS-specific. Check out this.

On Unix

On Unix, OsStr implements the std::os::unix::ffi:: OsStrExt trait, which augments it with two methods, from_bytes and as_bytes . These do inexpensive conversions from and to UTF-8 byte slices.

Additionally, on Unix OsString implements the std::os::unix::ffi:: OsStringExt trait, which provides from_vec and into_vec methods that consume their arguments, and take or produce vectors of u8 .

On Windows

On Windows, OsStr implements the std::os::windows::ffi:: OsStrExt trait, which provides an encode_wide method. This provides an iterator that can be collect ed into a vector of u16 .

Additionally, on Windows OsString implements the std::os::windows:ffi:: OsStringExt trait, which provides a from_wide method. The result of this method is an OsString which can be round-tripped to a Windows string losslessly.

[Yogi39]

Sorry, I forgot to put in I am only interested in Linux.

I started this reply by copying some of my efforts & the error messages. It seemed very verbose & pointless. I had studied Path, PathBuf, Ostr, std::convert etc. crates until I was going bonkers. I was aware that this conversion was explicitly or implicitly going to regard my simple 'String of bytes' as an Ostr at one point , but the syntax? (The number of Rust 'types' to represent what I referred to there as a simple 'String of bytes' I find overwhelming. AFAICS what I have and what I want are both just this.) I am sure I have driven myself to miss the obvious. So please could I trouble you to be more detailed from my original effort:-

let pth = PathBuf::from(&rdnew[60..]);

result:

trait std::convert::AsRef<std::ffi::OsStr> is not implemented for [u8]

(rdnew is my vector). How do I correct this without going round in circles?

(The code I am producing will run through about 5 million such vectors, but once it has successfully produced an answer it may never be used again. A lot of my programming is like this.)

Thanks a lot.
Yogi.

[mbrubeck]

use std::os::unix::ffi::OsStringExt;
let pth = PathBuf::from(OsString::from_vec(rdnew[60..].to_vec()));

or, equivalently:

use std::os::unix::ffi::OsStrExt;
let pth = Path::new(&OsStr::from_bytes(&rdnew[60..])).to_path_buf();

[Yogi39]

Thank you very much, both snippets seem from a syntax point of view. However, and worrying, the path does not seem to work, and I have no idea why. My code evolves around a set of nearly five million vectors each referring to a file. (Iteratively referred to in my code as rdnew. The path is in the vector from byte 60 on.) I hoped to write a twenty minute playground program to estimate the size of storage I would need if I decided to copy all the files referred to into one place. I only need the sum of the size of each file. I am finding it unexpectedly difficult to make rust do this. My present code snippet that fails is as follows:-

let mtdt=fs::symlink_metadata(PathBuf::from(OsString::from_vec(rdnew[60..].to_vec())));
if mtdt.is_err(){println!("fmtdt failed with {:?}\nrdnew[60..]= {}"
,mtdt,String::from_utf8_lossy(&rdnew[60..]));continue}//~2

(the //~2 tells me how 'deep' in terms of loops/scopes or whatever. In rust (only) it is always the no. left curly brackets minus right curly brackets in the code so far. Here ~1 = I am inside fn Main, ~2 = I am inside the loop iterating through the vectors. ~3 = I am now one level deep in an 'if' or whatever.)

To complete the context, the code ends from here simply:-

let mtdt=mtdt.unwrap();
if mtdt.is_file() {rqsz+=mtdt.len()+8192;} else {rqsz+=4096;} }//~1
println!("required size={}",rqsz); }//~0

The first snippet above fails every time (example chosen to be one where the path is all ASCSI):-

fmtdt failed with Err(Os { code: 2, kind: NotFound, message: "No such file or directory" })
rdnew[60..]= /media/richard/bkup02usb/hewhdb1/pictures/sonevxxx/sonev093/dsc89460.jpg

The path exists and is valid . To prove this I open a terminal, type ls -l and then copy/paste the path from the error message (I.e. from just after rdnew[60]=)

richard@yogi2 ~ $ ls -l /media/richard/bkup02usb/hewhdb1/pictures/sonevxxx/sonev093/dsc89460.jpg
-rwxrwxrwx 1 richard richard 2852309 Sep 3 2017 /media/richard/bkup02usb/hewhdb1/pictures/sonevxxx/sonev093/dsc89460.jpg
richard@yogi2 ~ $

Please can you (or anybody else) please tell me why my code does not find it and add 2852309 (with 8k more for overheads) to my variable rqsz?

[mbrubeck]

The first thing I would check is that the path doesn't contain any leading or trailing whitespace characters or non-printing control characters. Instead of printing it with the "{}" formatting string, try "{:?}" (Debug formatting), which will escape any control characters.

Also, just as an aside, you can pass an &OsStr directly to symlink_metadata, which would simplify this line:

let mtdt = fs::symlink_metadata(OsStr::from_bytes(&rdnew[60..]));