Your Name
2bc5b56ff9
erdtree (erd) replaces GNU tree as the tree-style directory lister. It uses a parallel walker (the `ignore` crate, one thread per core) where GNU tree is strictly serial. Benchmarked over 658k entries under ~/development, best-of-3, warm cache, plain output, hidden included, gitignore off: erd (16 threads) 2.35s GNU tree 2.1.0 4.22s (1.8x slower) erd (1 thread) 4.97s (so ~2.1x thread scaling) eza --tree 15.87s (6.8x slower) Three changes: - home_config/erdtree/.erdtree.toml — erd ships as a disk-usage tool, so its stock defaults are root-at-bottom, size-sorted, with a du column. The top-level table overrides that to behave like GNU tree; `erd -c du` keeps the original behaviour and `erd -c all` disables the gitignore/hidden filtering. Stowed to ~/.config/erdtree. - home_root/.zshrc — `alias tree='erd'`, plus ~/.zfunc prepended to fpath for the generated completion. The fpath line goes before the primary compinit at the top of the file rather than adding a third compinit call after it. - lk-installers/install-apps — erd is now a runtime assumption of the dotfiles (the alias depends on it), so a fresh box must get it: brew on macOS, cargo on Linux (not packaged in Debian apt). The zsh completion is generated per installed version rather than tracked, matching how the other generated completions in this repo are handled. Caveat worth knowing: erd respects .gitignore by default, which GNU tree does not. `-i` disables it, `-.` shows hidden files.
40 lines
1.6 KiB
TOML
40 lines
1.6 KiB
TOML
# erdtree (erd) configuration — /home/lkraven/.config/erdtree/.erdtree.toml
|
|
#
|
|
# The top-level table below is the DEFAULT for a bare `erd`. Named tables are
|
|
# opt-in via `erd -c <name>`.
|
|
#
|
|
# Rationale: erd ships as a disk-usage tool — its stock defaults are root-at-
|
|
# bottom, sorted by size, with a du column. That is a fine `du` replacement but
|
|
# a surprising `tree`. This config makes the bare invocation behave like GNU
|
|
# tree (root at top, name-sorted, dirs first, no size column) and moves the
|
|
# disk-usage behaviour behind `erd -c du`.
|
|
#
|
|
# Note: erd respects .gitignore by default. That is kept — it is why listing a
|
|
# repo is fast and readable. Pass `-i` to ignore .gitignore, `-.` for hidden
|
|
# files, `-. --no-git` for hidden-but-not-.git.
|
|
|
|
# ---- default: tree-like folder listing ----
|
|
layout = "inverted" # root at top, like GNU tree
|
|
sort = "name" # lexicographic, like GNU tree
|
|
dir-order = "first" # directories above files
|
|
suppress-size = true # no disk-usage column
|
|
no-progress = true # no progress indicator on fast listings
|
|
|
|
# ---- `erd -c du` : the disk-usage mode erd was actually built for ----
|
|
[du]
|
|
layout = "inverted"
|
|
sort = "rsize" # largest first
|
|
dir-order = "first"
|
|
human = true # human-readable byte sizes
|
|
disk-usage = "physical" # actual blocks on disk
|
|
|
|
# ---- `erd -c all` : everything, nothing hidden or ignored ----
|
|
[all]
|
|
layout = "inverted"
|
|
sort = "name"
|
|
dir-order = "first"
|
|
suppress-size = true
|
|
no-progress = true
|
|
hidden = true
|
|
no-ignore = true
|