Fix CSV path joining and relative exclusions; add scan benchmark script

This commit is contained in:
Andy Dixon
2026-10-02 08:37:18 +01:00
parent 8559616b65
commit d7e1d3e620
5 changed files with 75 additions and 1 deletions
+20
View File
@@ -21,6 +21,26 @@ bin/treegen --out /tmp/t --directories 50000 --files 1000000 --max-depth 20 --sp
bin/mapsize /tmp/t bin/mapsize /tmp/t
``` ```
## Benchmarks
```sh
make benchmark # Go micro-benchmarks (layout, render, filter, scan)
scripts/bench-scan.sh /path 1 2 4 8 # wall time and peak RSS per worker count
sudo DROP_CACHES=1 scripts/bench-scan.sh /path # cold-cache runs
MAPSIZE_MEM_ROOT=/path go test ./internal/scan -run TestMemoryPerNode -v
```
Paste results into PERFORMANCE.md with the machine description. Never edit
numbers by hand.
## Terminal restoration checks
```sh
MAPSIZE_DEBUG_PANIC=ui bin/mapsize . # panics on the first key press
MAPSIZE_DEBUG_PANIC=scanner bin/mapsize / # panics in a scanner goroutine
stty -a # echo/icanon must be back on
```
## Profiling ## Profiling
```sh ```sh
+2 -1
View File
@@ -8,6 +8,7 @@ import (
"encoding/json" "encoding/json"
"fmt" "fmt"
"io" "io"
"path/filepath"
"strconv" "strconv"
"time" "time"
@@ -158,7 +159,7 @@ func CSV(w io.Writer, t *inventory.Tree, root inventory.NodeID, maxDepth int) er
return nil return nil
} }
for _, c := range t.SortedChildren(id, inventory.SizeAllocated) { for _, c := range t.SortedChildren(id, inventory.SizeAllocated) {
if err := rec(c, path+"/"+t.Node(c).Name, depth-1); err != nil { if err := rec(c, filepath.Join(path, t.Node(c).Name), depth-1); err != nil {
return err return err
} }
} }
+11
View File
@@ -138,6 +138,17 @@ type inodeKey struct{ dev, ino uint64 }
// Start stats root, creates the tree and begins scanning in the background. // Start stats root, creates the tree and begins scanning in the background.
func Start(ctx context.Context, root string, opts Options) (*Scanner, error) { func Start(ctx context.Context, root string, opts Options) (*Scanner, error) {
opts.defaults() opts.defaults()
// Path-style exclusions are matched against absolute paths.
excl := make([]string, len(opts.Excludes))
for i, p := range opts.Excludes {
excl[i] = p
if strings.ContainsAny(p, `/\`) {
if a, err := filepath.Abs(p); err == nil {
excl[i] = a
}
}
}
opts.Excludes = excl
abs, err := filepath.Abs(root) abs, err := filepath.Abs(root)
if err != nil { if err != nil {
return nil, err return nil, err
+6
View File
@@ -237,6 +237,12 @@ func TestExcludes(t *testing.T) {
if tr.Stats.Excluded != 2 || tr.Node(0).Files != 1 { if tr.Stats.Excluded != 2 || tr.Node(0).Files != 1 {
t.Fatalf("excluded=%d files=%d", tr.Stats.Excluded, tr.Node(0).Files) t.Fatalf("excluded=%d files=%d", tr.Stats.Excluded, tr.Node(0).Files)
} }
// Relative path patterns resolve against the working directory.
t.Chdir(root)
tr = runScan(t, root, Options{Workers: 2, Excludes: []string{"./skip"}})
if tr.Stats.Excluded != 1 {
t.Fatalf("relative exclude: excluded=%d", tr.Stats.Excluded)
}
} }
func TestUnicodeAndHostileNames(t *testing.T) { func TestUnicodeAndHostileNames(t *testing.T) {
+36
View File
@@ -0,0 +1,36 @@
#!/usr/bin/env bash
# Records scan benchmarks as Markdown table rows for PERFORMANCE.md.
#
# scripts/bench-scan.sh PATH [workers...] (default workers: 1 2 4 8 16)
#
# Linux only (uses /usr/bin/time, df, lsblk). Runs each configuration three
# times against a warm cache; to measure a cold cache, run as root with
# DROP_CACHES=1, which drops the page cache before every run.
set -euo pipefail
path=${1:?usage: bench-scan.sh PATH [workers...]}
shift || true
workers=${*:-1 2 4 8 16}
bin=${MAPSIZE:-bin/mapsize}
[[ -x $bin ]] || make build >/dev/null
fs=$(df -T "$path" | awk 'NR==2 {print $2}')
dev=$(df "$path" | awk 'NR==2 {print $1}')
medium=$(lsblk -no ROTA "$dev" 2>/dev/null | head -1 | sed 's/1/rotational/;s/0/solid-state/')
counts=$("$bin" --no-ui "$path" 2>/dev/null | awk '/^Files/ {f=$2} /^Directories/ {d=$2} END {print f" files, "d" dirs"}')
echo "<!-- $(date -u +%FT%TZ) $(uname -srm), $(nproc) CPUs -->"
echo "Path: \`$path\` · $fs on ${medium:-unknown medium} · $counts"
echo
echo "| workers | cache | wall (s) | peak RSS (MB) |"
echo "|---:|---|---:|---:|"
for w in $workers; do
for i in 1 2 3; do
cache=warm
if [[ ${DROP_CACHES:-} == 1 ]]; then
sync; echo 3 > /proc/sys/vm/drop_caches; cache=cold
fi
out=$( { /usr/bin/time -f "%e %M" "$bin" --no-ui --workers "$w" "$path" >/dev/null; } 2>&1 | tail -1)
read -r secs kb <<<"$out"
echo "| $w | $cache | $secs | $((kb / 1024)) |"
done
done