xdebug-mcp/internal/cachegrind/parser.go

package cachegrind

import (
	"bufio"
	"compress/gzip"
	"fmt"
	"io"
	"os"
	"strconv"
	"strings"
)

// ParseFile opens a cachegrind file (plain text or gzip) and returns a Profile.
func ParseFile(path string) (*Profile, error) {
	f, err := os.Open(path)
	if err != nil {
		return nil, fmt.Errorf("open %s: %w", path, err)
	}
	defer f.Close()

	var r io.Reader = f
	if strings.HasSuffix(path, ".gz") {
		gr, gzErr := gzip.NewReader(f)
		if gzErr != nil {
			// fall back to plain text
			if _, seekErr := f.Seek(0, io.SeekStart); seekErr != nil {
				return nil, fmt.Errorf("gzip open %s: %w", path, gzErr)
			}
		} else {
			defer gr.Close()
			r = gr
		}
	}

	return Parse(r)
}

// Parse reads cachegrind format from r and returns a Profile.
// All cost lines for a function are summed into inclusive costs (includes callees).
// Two passes are used: the first builds the function table, the second resolves call edges.
func Parse(r io.Reader) (*Profile, error) {
	scanner := bufio.NewScanner(r)
	scanner.Buffer(make([]byte, 4*1024*1024), 4*1024*1024)

	var lines []string
	for scanner.Scan() {
		lines = append(lines, scanner.Text())
	}
	if err := scanner.Err(); err != nil {
		return nil, err
	}

	p := &Profile{ByName: make(map[string][]*Function)}

	// Pass 1: collect events count and build fnTable (alias → *Function).
	flTable := map[int]string{}
	fnTable := map[int]*Function{}
	nEvents := 0
	currentFile1 := ""

	for _, line := range lines {
		switch {
		case strings.HasPrefix(line, "events:"):
			p.Events = strings.Fields(strings.TrimPrefix(line, "events:"))
			nEvents = len(p.Events)
		case strings.HasPrefix(line, "cmd:"):
			p.Cmd = strings.TrimSpace(strings.TrimPrefix(line, "cmd:"))
		case strings.HasPrefix(line, "fl="):
			alias, name := parseAlias(line[3:])
			if name != "" {
				flTable[alias] = name
			}
			currentFile1 = flTable[alias]
		case strings.HasPrefix(line, "fn="):
			alias, name := parseAlias(line[3:])
			if name != "" {
				fn := &Function{
					Name:  name,
					File:  currentFile1,
					Costs: make([]int64, nEvents),
				}
				fnTable[alias] = fn
				p.Functions = append(p.Functions, fn)
				p.ByName[name] = append(p.ByName[name], fn)
			}
		}
	}

	// Pass 2: accumulate costs and resolve call edges (all functions now known).
	var currentFn *Function
	var pendingCallee *Function
	pendingCallCount := int64(0)
	inCallCtx := false

	for _, line := range lines {
		if line == "" || strings.HasPrefix(line, "#") {
			continue
		}

		switch {
		case strings.HasPrefix(line, "events:"), strings.HasPrefix(line, "cmd:"),
			strings.HasPrefix(line, "version:"), strings.HasPrefix(line, "creator:"),
			strings.HasPrefix(line, "part:"), strings.HasPrefix(line, "positions:"),
			strings.HasPrefix(line, "ob="), strings.HasPrefix(line, "totals:"),
			strings.HasPrefix(line, "summary:"):
			// already handled or irrelevant

		case strings.HasPrefix(line, "fl="):
			inCallCtx = false

		case strings.HasPrefix(line, "fn="):
			alias, _ := parseAlias(line[3:])
			currentFn = fnTable[alias]
			inCallCtx = false

		case strings.HasPrefix(line, "cfl="):
			// called file alias — not needed

		case strings.HasPrefix(line, "cfn="):
			alias, _ := parseAlias(line[4:])
			pendingCallee = fnTable[alias]

		case strings.HasPrefix(line, "calls="):
			parts := strings.Fields(line[6:])
			if len(parts) > 0 {
				pendingCallCount, _ = strconv.ParseInt(parts[0], 10, 64)
			}
			inCallCtx = true

		default:
			if len(line) == 0 || line[0] < '0' || line[0] > '9' {
				continue
			}
			costs, err := parseCostLine(line, nEvents)
			if err != nil {
				continue
			}
			if inCallCtx && currentFn != nil && pendingCallee != nil {
				call := &Call{
					Caller: currentFn,
					Callee: pendingCallee,
					Count:  pendingCallCount,
					Costs:  costs,
				}
				currentFn.Calls = append(currentFn.Calls, call)
				pendingCallee.CalledBy = append(pendingCallee.CalledBy, call)
				pendingCallee = nil
				pendingCallCount = 0
				inCallCtx = false
			}
			if currentFn != nil {
				addCosts(currentFn.Costs, costs)
			}
		}
	}

	return p, nil
}

// parseAlias parses "(N)" or "(N) name" from the value portion of fl=/fn=.
func parseAlias(s string) (int, string) {
	if !strings.HasPrefix(s, "(") {
		return 0, ""
	}
	end := strings.Index(s, ")")
	if end < 0 {
		return 0, ""
	}
	alias, _ := strconv.Atoi(s[1:end])
	name := ""
	if len(s) > end+1 {
		name = strings.TrimSpace(s[end+1:])
	}
	return alias, name
}

// parseCostLine parses "linenum cost1 cost2 ..." and returns the cost slice.
func parseCostLine(line string, nEvents int) ([]int64, error) {
	parts := strings.Fields(line)
	if len(parts) < 2 {
		return nil, fmt.Errorf("short cost line: %q", line)
	}
	costs := make([]int64, nEvents)
	for i := 0; i < nEvents && i+1 < len(parts); i++ {
		v, err := strconv.ParseInt(parts[i+1], 10, 64)
		if err != nil {
			return nil, err
		}
		costs[i] = v
	}
	return costs, nil
}

func addCosts(dst, src []int64) {
	for i := range src {
		if i < len(dst) {
			dst[i] += src[i]
		}
	}
}