Add parse_glob module and update du to use parse_glob
#3754
Conversation
|
Any tips on how to debug the new test failure in I did on my machine (linux) and it was fine. |
|
@ackerleytng That's very curious. Maybe you could download the logs from the CI and look at those? If it's a CI only issue we might have to ignore it or fix the CI somehow. Edit: Oh I think that test just has some intermittent failures. I'll rerun it. |
|
I diffed the this test run with an earlier working one and nothing stood out, other than This test
Do we know if these tests are sufficiently independent of each other? |
|
I just noticed this: https:/uutils/coreutils/blob/main/util/build-gnu.sh#L171 Since this test uses more than one process and is timing dependent, is it something to do with the hardware (performance) of the CI machine? |
|
Yeah that looks like something we should fix outside this PR |
|
Feel free to merge this if you're okay with fixing this separately! |
|
Yeah!
|
| let mut inodes: HashSet<FileInfo> = HashSet::new(); | ||
| if let Some(inode) = stat.inode { | ||
| inodes.insert(inode); | ||
| if let Ok(stat) = Stat::new(path, &options) { |
There was a problem hiding this comment.
Could you please add a comment explaining what this block is doing? It isn't super obvious reading the code
There was a problem hiding this comment.
Please see newest changes!
| @@ -745,6 +745,27 @@ fn test_du_exclude_mix() { | |||
| assert!(!result.stdout_str().contains("qzerty")); | |||
| assert!(!result.stdout_str().contains("azeaze")); | |||
| assert!(result.stdout_str().contains("xcwww")); | |||
|
|
|||
| // Negation in glob should work with both ^ and ! | |||
| let result = ts | |||
There was a problem hiding this comment.
maybe create a new test for this?
ackerleytng
force-pushed
the
main
branch
3 times, most recently
from
de02ebb
to
6aa9aa1
Compare
Fixes #3628
The
parse_globmodule uses the same logic implemented inrust-lang/glob: https:/rust-lang/glob/blob/master/src/lib.rs#L613 to identify negation in globs. When this pattern is identified, the^in the pattern is replaced with!and glob'sPatternis built from the string.A disadvantage of this approach is that inconsistencies may result, if this pattern identification method diverges, when rust-lang/glob's parsing of character set negation changes.
However, I think this approach avoids having to maintain or rebuild glob parsing logic, and is the best option I could think of given that
rust-lang/glob's implementation is not open to extension to support multiple character set negation markers.With the
parse_globmodule, fixingduis just a matter of using the new module.GNU
lsanddircolorsalso usefnmatch. Those two utilities will be updated to useparse_globin separate PRs.