Janady13/missionmatch / cmd/opportunity-extract/oom_diag_test.go

Source identity

Field	Recorded value
Source type	GitHub source
Original record	Janady13/missionmatch / cmd/opportunity-extract/oom_diag_test.go
Date	Not specified in source
Repository	Janady13/missionmatch
Revision	720d47fae476b0d38477ae4401697d372b1ea2bc
Branch	feature/opportunity-to-outcome
Path or archive member	cmd/opportunity-extract/oom_diag_test.go
Original SHA-256	00b7c09a7a25809554790430fb8d3a5ec56ed200ba67a67f0004663de71ce785
Git blob SHA	02900144200fd64d95d460fed7dd7c0e80da90a1
Reading-copy SHA-256	00b7c09a7a25809554790430fb8d3a5ec56ed200ba67a67f0004663de71ce785

Original source ↗

Recovered source transcript

Download complete source text ↓56 lines · 1.5 KB

package main

import (
	"bytes"
	"os"
	"runtime"
	"testing"

	pdflib "github.com/ledongthuc/pdf"
)

// A DIAGNOSTIC, not an assertion. It answers one question about the document that OOM killed the extractor:
// does memory accumulate across pages, which a bound can fix, or does one page explode inside the library,
// which a bound cannot. The answer decides the shape of the fix.
func TestTE6DiagnoseWhereTheMemoryGoes(t *testing.T) {
	path := os.Getenv("MM_TEST_TE6_PDF")
	if path == "" {
		t.Skip("set MM_TEST_TE6_PDF to run")
	}
	raw, err := os.ReadFile(path)
	if err != nil {
		t.Skipf("fixture unreadable: %v", err)
	}
	heapMB := func() uint64 {
		var m runtime.MemStats
		runtime.ReadMemStats(&m)
		return m.HeapAlloc / (1 << 20)
	}
	defer func() {
		if rec := recover(); rec != nil {
			t.Logf("panicked: %v", rec)
		}
	}()

	t.Logf("heap before NewReader: %d MB", heapMB())
	rd, err := pdflib.NewReader(bytes.NewReader(raw), int64(len(raw)))
	if err != nil {
		t.Fatalf("NewReader: %v", err)
	}
	t.Logf("heap after NewReader: %d MB, NumPage=%d", heapMB(), rd.NumPage())

	for i := 1; i <= rd.NumPage() && i <= 12; i++ {
		p := rd.Page(i)
		if p.V.IsNull() {
			t.Logf("page %d null, heap %d MB", i, heapMB())
			continue
		}
		txt, e := p.GetPlainText(nil)
		t.Logf("page %d: text=%d bytes err=%v heap=%d MB", i, len(txt), e, heapMB())
		if heapMB() > 400 {
			t.Logf("STOP: heap passed 400 MB at page %d, so growth is PER PAGE and a bound can fix it", i)
			return
		}
	}
}