test(ingestion): behavioural race tests for raw/ reads; tighten the lint
CI / Lint / Test / Vet (pull_request) Successful in 36s
CI / Mirror to GitHub (pull_request) Skipped

PR #156 round 3: a readRawNote rewrite to os.ReadFile or Root.ReadFile
passed the lint and brought back the escape (19k leaks in 4s) and the
FIFO hang. Listing forbidden APIs is whack-a-mole, so the control now
rests on behaviour:

- TestRawReadersLeakNothingUnderASymlinkSwapRace and
  TestRawReadersDoNotHangUnderAFIFOSwapRace swap a regular note, a
  symlink to an outside file, or a FIFO into raw/r.md while ListPending
  and ReadIngestFile loop for 1.5s; any leaked byte or a hang fails.
- TestProcessDir_DoesNotHangUnderAFIFOSwapRace does the same for the
  unattended watcher (read and archive copy).

Flipped: os.ReadFile -> leak + hang tests red; Root.ReadFile -> hang
test red; copyFromRoot back to Root.Open -> watcher test red.

The lint stays as a second layer: it also forbids ReadFile, matches the
selector (so a stored method value is caught), refuses a dot-import of
extract, and says it only covers the three files it parses.

Refs #153

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-10-03 17:07:58 +02:00
co-authored by Claude Opus 5.5
parent 809420070b
commit d5b2be0ecf
3 changed files with 179 additions and 6 deletions
+112
View File
@@ -0,0 +1,112 @@
package api_test
import (
"os"
"path/filepath"
"strings"
"sync/atomic"
"syscall"
"testing"
"time"
"github.com/mathiasbq/hyperguild/ingestion/internal/api"
)
// hyperguild#153 / PR #156: behavioural race tests, so the control does not
// rest on lints that list forbidden APIs (round 3: an os.ReadFile or
// Root.ReadFile rewrite passed the lint and leaked 19k times in 4s). A
// swapper keeps renaming a regular note, a symlink to an outside file, or a
// FIFO into raw/r.md while the public readers loop. Any leaked byte fails;
// any hang fails.
const raceSentinel = "SENTINEL-153-RACE"
func swap(raw, outside string, fifo bool, stop *atomic.Bool) {
tgt := filepath.Join(raw, "r.md")
for i := 0; !stop.Load(); i++ {
tmp := filepath.Join(raw, ".swap")
_ = os.Remove(tmp)
switch {
case i%2 == 0:
_ = os.WriteFile(tmp, []byte("# r\n\ninside"), 0o644)
case fifo:
_ = syscall.Mkfifo(tmp, 0o644)
default:
_ = os.Symlink(outside, tmp)
}
_ = os.Rename(tmp, tgt)
}
}
func raceBrain(t *testing.T) (brain, raw, outside string) {
t.Helper()
brain = t.TempDir()
raw = filepath.Join(brain, "raw")
if err := os.MkdirAll(raw, 0o755); err != nil {
t.Fatal(err)
}
outside = filepath.Join(t.TempDir(), "secret.md")
if err := os.WriteFile(outside, []byte("# s\n\n"+raceSentinel), 0o644); err != nil {
t.Fatal(err)
}
if err := os.WriteFile(filepath.Join(raw, "r.md"), []byte("# r\n\ninside"), 0o644); err != nil {
t.Fatal(err)
}
return brain, raw, outside
}
// readers runs every public raw/ reader once and returns all text they
// produced (excerpts, extracted text).
func readers(brain, raw string) string {
var b strings.Builder
if ns, err := api.ListPending(brain); err == nil {
for _, n := range ns {
b.WriteString(n.Excerpt)
}
}
if s, err := api.ReadIngestFile(brain, filepath.Join(raw, "r.md")); err == nil {
b.WriteString(s)
}
return b.String()
}
func TestRawReadersLeakNothingUnderASymlinkSwapRace(t *testing.T) {
brain, raw, outside := raceBrain(t)
var stop atomic.Bool
go swap(raw, outside, false, &stop)
defer stop.Store(true)
leaks, iters := 0, 0
for deadline := time.Now().Add(1500 * time.Millisecond); time.Now().Before(deadline); iters++ {
leaks += strings.Count(readers(brain, raw), raceSentinel)
}
if leaks > 0 {
t.Fatalf("outside file leaked %d times in %d iterations", leaks, iters)
}
t.Logf("%d iterations, 0 leaks", iters)
}
func TestRawReadersDoNotHangUnderAFIFOSwapRace(t *testing.T) {
brain, raw, outside := raceBrain(t)
var stop atomic.Bool
go swap(raw, outside, true, &stop)
var iters atomic.Int64
done := make(chan struct{})
go func() {
for deadline := time.Now().Add(1500 * time.Millisecond); time.Now().Before(deadline); {
_ = readers(brain, raw)
iters.Add(1)
}
close(done)
}()
select {
case <-done:
stop.Store(true)
case <-time.After(6 * time.Second):
stop.Store(true)
// Unblock the stuck open so the test process can exit.
if f, err := os.OpenFile(filepath.Join(raw, "r.md"), os.O_WRONLY|syscall.O_NONBLOCK, 0); err == nil {
_ = f.Close()
}
t.Fatalf("a raw/ reader hung on a swapped-in FIFO after %d iterations", iters.Load())
}
}