Add support for GNU-compatible printf

[refi64]

This adds initial support for GNU find-compatible printf syntax. As far
as I can tell, it should implement all the specifiers, with the
following exceptions:

• Using zero as padding instead of space: GNU find isn't consistent with
  which numeric format specifiers support this.
• Specifying precision (%x.yz): I believe sparseness (%S)
  and seconds-since-epoch (%A@ / %C@ / %T@) are the only ones that
  allow this, making it only partially useful.
• Printing SELinux contexts (%Z): requires some extra dependencies and
  isn't within our scope.

uucore is added as a dependency, because it has several utility
functions that we need, the complexity of which is high enough that
using the extra dependency is quite helpful.

This also stubs out many of the specifiers for Windows, simply because I
have no idea how these should work there. (The defaults chosen should be
"good enough" for the moment, though.)

Signed-off-by: Ryan Gonzalez ryan.gonzalez@collabora.com

---

Okay, so some more meta notes on this:

• Right now, printing the filesystem doesn't work reliably without: uucore::fsext: Fix mountinfo parsing w/ multiple optional fields coreutils#2778
• I'm not sure if what I put for the copyright in the new file is correct, please point it out if not!
• Also not sure if it would be worth removing printer.rs in favor of just putting Printf::new("%p").unwrap() (which I believe should be identical) in the main file.

[codecov]

Codecov Report

Merging #120 (dd443b9) into master (d385362) will increase coverage by 2.10%.
The diff coverage is 65.06%.

@@            Coverage Diff             @@
##           master     #120      +/-   ##
==========================================
+ Coverage   47.70%   49.80%   +2.10%     
==========================================
  Files          18       19       +1     
  Lines        2111     2594     +483     
  Branches      495      702     +207     
==========================================
+ Hits         1007     1292     +285     
- Misses        933     1046     +113     
- Partials      171      256      +85     

Impacted Files	Coverage Δ
src/find/matchers/delete.rs	79.16% <ø> (ø)
src/find/mod.rs	61.36% <ø> (ø)
tests/find_cmd_tests.rs	70.58% <47.36%> (-18.78%)	⬇️
src/find/matchers/mod.rs	65.78% <50.00%> (-0.43%)	⬇️
src/find/matchers/printf.rs	67.70% <67.70%> (ø)
src/lib.rs	25.54% <0.00%> (+2.81%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d385362...dd443b9. Read the comment docs.

[wlozano0collabora]

Very strange test failure, it doesn't seem to be related to this PR.

[refi64]

oh oops I had fixed it locally but forgot to push, just did it now

[sylvestre]

Would it be possible to have high level tests?
https://github.com/uutils/findutils/blob/master/tests/find_cmd_tests.rs
thanks

[refi64]

@sylvestre I added some high-level tests in src/find/mod.rs with a lot of the others there, does that look okay?

[sylvestre]

@sylvestre I added some high-level tests in src/find/mod.rs with a lot of the others there, does that look okay?

could you please move them into the tests directory? I prefer to have all integration tests at the same place. thanks

[refi64]

@sylvestre Ah I misunderstood, just moved it over and actually testing the find binary itself (not find_main).

[refi64]

Okay got CI on Windows working again with the moved tests 😅 Should be good now. (I also pointed it to uucore from Git so that it has the mentioned fix for grabbing the filesystem (otherwise %F doesn't work).)

[sylvestre]

thanks
Seems that the coverage could be improved:
https://codecov.io/github/uutils/findutils/commit/f9ab5847cf4d41bfc15ed9464218ec6915977641
do you think you could add a few more tests ?

[refi64]

@sylvestre I added a few more tests, with that, though, most of the still-remaining red is things that can't easily be tested in CI due to a high dependence on host system config (e.g. filesystem type, block count, sparseness)

[refi64]

(and CI is passing now, again)

[sylvestre]

fails on ubuntu:

---- find::matchers::printf::tests::test_printf_special_types stdout ----
thread 'find::matchers::printf::tests::test_printf_special_types' panicked at 'called `Result::unwrap()` on an `Err` value: Error { depth: 2, inner: Io { path: Some("/dev/shm/multipath"), err: Os { code: 13, kind: PermissionDenied, message: "Permission denied" } } }', src/find/matchers/mod.rs:433:47
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[refi64]

@sylvestre Well in retrospect, trying to use /dev here was a terrible idea 😅 new version uses some custom FIFO/socket files instead in order to test special file types

[sylvestre]

thanks and sorry for the latency :(