When running:
dust dir_a dir_b
if a file was hard linked in both dir_a and dir_b it would be double
counted.
This fix resolves this by keeping the shared hashmap around between runs
for the second and subsequent arguments.
https://github.com/bootandy/dust/issues/282
- Try to use iterator adapters and collect in various places, where possible. This especially benefits draw_it.
- Try to use `.map` and other similar methods on Options and Results, where possible
- Replaced nearly all clones with reference-based equivalents
- Summarizing nodes by file extension is now much more efficient
- PartialOrd and PartialEq implementations now agree
- Replace #[cfg(...)] function definitions with simpler if cfg!(...) equivelents
- Simplify CLI Values handling by taking advantage of Values::default
- Various spelling corrections in comments
- Add `ColorState` enum to replace bool, for clarity
- Fix tests that break under some detected terminal widths when paths are long
- Use sort_by instead of (sort, reverse)
- Use new `ExtensionNode` struct internally to simplify extension aggregation code
Before we assumed 5 chars for printing the size of a node. This change
allows us to check if we need 5 chars or if we can manage with 4 which
may save us a character.
(1023B = 5 chars, 1.2K = 4 chars)
Note: Mac test runners create very long filenames in tmp directories. To
fix this we must either
1) Tell the terminal it has a very wide width
2) Duplicate the code to reduce the size of the filename (add .. at the
end).
This commit changes from (2) to (1)
Now files are iterated thru in a preditermined order we can tighten the
boundaries of the test_hard_sym_link test.
Link will not be displayed, only the file name
The code calculating the width of a row to use vs the width of the
terminal wasn't quite right.
If we had 2 long filepaths one at the top of the dir tree and one at the
bottom the old code would shorten both equally. This doesn't make sense
as the one at the shallowest part of the tree would have used less
screen real estate and so should show a longer part of the filepath.
Code changes:
Removed ignore & channel crates. Using a single reciever thread to build
a hashmap to prevend duplicate inodes being reported gave a severe
performance penalty
Using rayon crate with some hand crafted file traversal has improved
performance aprox 10X
Behaviour changes:
Removed parameter 'limit by filesystem' - don't think this is used, and
I only added it as it was easy to add with the ignore crate.
Sym links will now not appear in the output tree unless using '-s'
'apparent-size' flag
Change behaviour of multiple args so that it unifies them and
compares them under one tree instead of treating them
individually: https://github.com/bootandy/dust/issues/136
https://github.com/bootandy/dust/issues/92
Not 100% sure if this code is clean yet, may well be a better way to do
it
Also:
Added test directory with incredably long name as test case.
Update test_symlinks.py for mac runners. Mac test runners create files
with very long names, hence the tests fail periodically unless they look
for a truncated name with '..' at the end.
Use the whole width of the terminal assume width of 80 if none found
Show percentages in the final column. Show ASCII bars indicating usage
of the underlying directories in the space inbetween.
Display (height of terminal - 10) entries by default.
Reverse the output order so largest is at the bottom.
Break up tests. Change older tests to check real output of program.
andy.boot
Nathan West