add new option: --output-format this controls how the output is
summarised and takes:
nothing = default behaviour
si = SI units (same as old --si flag)
b = bytes
kb = kb
kib = si kb
mb = mb
mib = si mb
....etc
Provide "No such file or directory" error if file is not found.
Provide "Unknown Error" if other error found
Should reduce confusion from the generic other error
If a user reports a crash from stacksize - we can try different stack
memory sizes to try and figure out the cause.
Do not try and initialise rayon on 32 bit builds
Fix rayon initialisation, may not have been called before due to badly
formed closure
- If `--depth` flag is not defined (or it has an invalid value), a value
from the config file will be used.
- If no `depth` entry in the config file (or there is no config file),
the default `usize::MAX` will be used.
Added test cases:
- no config and no flag defined -> `usize::MAX` should be used
- config defined, but flag is not defined -> config value should be used
- config is not defined, but flag is defined -> flag value should be used
- both config and flag is defined -> flag value should be used
Additional changes:
- Fixed some clippy issues.
- Added comments to the example `config.toml` file.
(copy from flag description)
Closes: #314
Signed-off-by: Balazs Nadasdi <efertone@pm.me>
Reduce & simlpify.
Cut feature: where we only report file count, it is always
useful to see file size as well.
cut feature: plural of file. If dust has only indexed a single file by
the time it prints the status then something has gone wrong, safe to
always assume a plural.
* v1.0
* renamed operations to be more clear
* put info later because there is still operations par of the preparing process
* updated the last line clearing
* changed name of module and structs to ones that make more sens
* Disable size computation when file_count option is set
* added sleep during the thread waiting
* use 1024 powered instead of 10 to compute showed number
* include DS_Store
* added files directories skipped information
* small format update
* implement the -H option
* put wait back
* remove PAtomicInfo since it's not used
* cargo fmt
* wrapped atomic operations to reduce overhead
* updated comments
* Use AtomicU64Wrapper instead of AtomicU64 in TotalSize
* update size suffix
* sto dividing size when larger than terabytes
* Fix use_iso flag not be set properly
* update properties display
* some reformating
* use stdout instead of print
* Moved config instance into main because it's easier to read
* merge base formatting into macro
* update name to be more intuitive and separated math operations for more flexibility
* print currently indexed path
* cargo fmt
* reset size between each target dirs
* Access to TotalSize rather than it's inner
* small comment change
* Update sysinfo version to 0.26.7
* fix: update use of sysinfo.system
System is now much quicker to start but requires an explicit call
to refresh memory else it deafults to 0 (oops)
* clippy: Fix new clippy
* fix: bug where hard links could be double counted
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
* Fix: depth=0 bug for multiple arguments
https://github.com/bootandy/dust/issues/282
* refactor filter.rs
* refactor filter.rs
* refactor create AggregateData for filter.rs
* feature: Support for dereference links -L follow
du has -L flag which allows it to dereference or follow
symlinks. Clone this feature into dust.
https://github.com/bootandy/dust/issues/276
* refactor dir_walker
I find this layout cleaner
* v1.0
* changed name of module and structs to ones that make more sens
* Disable size computation when file_count option is set
* added files directories skipped information
* implement the -H option
* wrapped atomic operations to reduce overhead
* used human_readable_number function in display module rather than our own
* implemented progress disabling
* cargo fmt & cargo clippy
Co-authored-by: Guillaume Gomez <guillaume1.gomez@gmail.com>
Co-authored-by: andy.boot <bootandy@gmail.com>
Small boxes do not have enough memory to create a large stack
Conversely we want a large stack size for large boxes with a very highly
nested directory structure.
Version: New version
Dislike the idea of passing a string into build_cli. By removing
a call to default_value we can side-step the problem.
Downside is we lose the error log if a user provides a bad depth
- 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
Low stack size could result in a stack overflow exception if traversing
very highly nested directories.
This makes core-dumps less likely but will not completely remove the
risk
example for creating directories:
mkdir -p $(for i in {1..5000}; do echo -n "qwe/"; done)
original issue:
https://github.com/bootandy/dust/issues/197
Allow -n for number_of_lines to be used with -d 'max depth'
Remove depth specific functions, the job is now handled by the mainline
Add depth as a field onto the Node object.
Fukushima Shogo
andy.boot
n4n5
zhaotao1
Efertone
NovaliX
Guillaume Gomez
DIRE
Collin Styles
Jedsek
Nathan West