zip: use []byte instead of string to prevent allocs (#537)

goos: linux goarch: amd64 pkg: github.com/gabriel-vasile/mimetype cpu: Intel(R) Core(TM) i7-10510U CPU @ 1.80GHz │ master │ dev │ │ sec/op │ sec/op vs base │ Common/.xlsx-8 482.0n ± 4% 446.7n ± 3% -7.33% (p=0.000 n=10) Common/.pptx-8 1.999µ ± 3% 1.542µ ± 3% -22.89% (p=0.000 n=10) Common/.docx-8 1.501µ ± 2% 1.139µ ± 3% -24.13% (p=0.000 n=10) Common/.tar-8 1.134µ ± 3% 1.127µ ± 2% ~ (p=0.670 n=10) Common/.zip-8 882.4n ± 1% 778.6n ± 4% -11.75% (p=0.000 n=10) Common/.pdf-8 363.9n ± 2% 352.9n ± 2% -3.02% (p=0.005 n=10) Common/.jpg-8 508.9n ± 4% 504.3n ± 2% ~ (p=0.190 n=10) Common/.png-8 497.1n ± 3% 495.2n ± 2% ~ (p=0.971 n=10) Common/.gif-8 594.9n ± 5% 591.8n ± 2% ~ (p=0.481 n=10) Common/.xls-8 602.7n ± 4% 598.2n ± 11% ~ (p=0.436 n=10) Common/.webm-8 1.541µ ± 3% 1.549µ ± 3% ~ (p=0.645 n=10) Common/.csv-8 10.89µ ± 1% 10.96µ ± 1% ~ (p=0.225 n=10) geomean 981.7n 919.3n -6.36% │ master │ dev │ │ B/op │ B/op vs base │ Common/.xlsx-8 312.0 ± 0% 288.0 ± 0% -7.69% (p=0.000 n=10) Common/.pptx-8 592.0 ± 0% 288.0 ± 0% -51.35% (p=0.000 n=10) Common/.docx-8 504.0 ± 0% 288.0 ± 0% -42.86% (p=0.000 n=10) Common/.tar-8 200.0 ± 0% 200.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.zip-8 224.0 ± 0% 192.0 ± 0% -14.29% (p=0.000 n=10) Common/.pdf-8 192.0 ± 0% 192.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.jpg-8 192.0 ± 0% 192.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.png-8 192.0 ± 0% 192.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.gif-8 192.0 ± 0% 192.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.xls-8 288.0 ± 0% 288.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.webm-8 192.0 ± 0% 192.0 ± 0% ~ (p=1.000 n=10) ¹ Common/.csv-8 7.331Ki ± 0% 7.332Ki ± 0% ~ (p=0.179 n=10) geomean 339.6 299.3 -11.85% ¹ all samples are equal │ master │ dev │ │ allocs/op │ allocs/op vs base │ Common/.xlsx-8 4.000 ± 0% 3.000 ± 0% -25.00% (p=0.000 n=10) Common/.pptx-8 15.000 ± 0% 3.000 ± 0% -80.00% (p=0.000 n=10) Common/.docx-8 13.000 ± 0% 3.000 ± 0% -76.92% (p=0.000 n=10) Common/.tar-8 3.000 ± 0% 3.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.zip-8 6.000 ± 0% 2.000 ± 0% -66.67% (p=0.000 n=10) Common/.pdf-8 2.000 ± 0% 2.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.jpg-8 2.000 ± 0% 2.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.png-8 2.000 ± 0% 2.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.gif-8 2.000 ± 0% 2.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.xls-8 3.000 ± 0% 3.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.webm-8 2.000 ± 0% 2.000 ± 0% ~ (p=1.000 n=10) ¹ Common/.csv-8 33.00 ± 0% 33.00 ± 0% ~ (p=1.000 n=10) ¹ geomean 4.339 2.991 -31.05% ¹ all samples are equal
gabriel-vasile · Jun 2, 2024 · cdceff9 · cdceff9
1 parent 77e3848
commit cdceff9
Show file tree

Hide file tree

Showing 2 changed files with 41 additions and 42 deletions.
diff --git a/internal/magic/ms_office.go b/internal/magic/ms_office.go
@@ -6,41 +6,41 @@ import (
 )
 
 var (
- xlsxSigFiles = []string{
- "xl/worksheets/",
- "xl/drawings/",
- "xl/theme/",
- "xl/_rels/",
- "xl/styles.xml",
- "xl/workbook.xml",
- "xl/sharedStrings.xml",
- }
- docxSigFiles = []string{
- "word/media/",
- "word/_rels/document.xml.rels",
- "word/document.xml",
- "word/styles.xml",
- "word/fontTable.xml",
- "word/settings.xml",
- "word/numbering.xml",
- "word/header",
- "word/footer",
- }
- pptxSigFiles = []string{
- "ppt/slides/",
- "ppt/media/",
- "ppt/slideLayouts/",
- "ppt/theme/",
- "ppt/slideMasters/",
- "ppt/tags/",
- "ppt/notesMasters/",
- "ppt/_rels/",
- "ppt/handoutMasters/",
- "ppt/notesSlides/",
- "ppt/presentation.xml",
- "ppt/tableStyles.xml",
- "ppt/presProps.xml",
- "ppt/viewProps.xml",
+ xlsxSigFiles = [][]byte{
+ []byte("xl/worksheets/"),
+ []byte("xl/drawings/"),
+ []byte("xl/theme/"),
+ []byte("xl/_rels/"),
+ []byte("xl/styles.xml"),
+ []byte("xl/workbook.xml"),
+ []byte("xl/sharedStrings.xml"),
+ }
+ docxSigFiles = [][]byte{
+ []byte("word/media/"),
+ []byte("word/_rels/document.xml.rels"),
+ []byte("word/document.xml"),
+ []byte("word/styles.xml"),
+ []byte("word/fontTable.xml"),
+ []byte("word/settings.xml"),
+ []byte("word/numbering.xml"),
+ []byte("word/header"),
+ []byte("word/footer"),
+ }
+ pptxSigFiles = [][]byte{
+ []byte("ppt/slides/"),
+ []byte("ppt/media/"),
+ []byte("ppt/slideLayouts/"),
+ []byte("ppt/theme/"),
+ []byte("ppt/slideMasters/"),
+ []byte("ppt/tags/"),
+ []byte("ppt/notesMasters/"),
+ []byte("ppt/_rels/"),
+ []byte("ppt/handoutMasters/"),
+ []byte("ppt/notesSlides/"),
+ []byte("ppt/presentation.xml"),
+ []byte("ppt/tableStyles.xml"),
+ []byte("ppt/presProps.xml"),
+ []byte("ppt/viewProps.xml"),
  }
 )
 

diff --git a/internal/magic/zip.go b/internal/magic/zip.go
@@ -3,7 +3,6 @@ package magic
 import (
  "bytes"
  "encoding/binary"
- "strings"
 )
 
 var (
@@ -43,7 +42,7 @@ func Zip(raw []byte, limit uint32) bool {
 
 // Jar matches a Java archive file.
 func Jar(raw []byte, limit uint32) bool {
- return zipContains(raw, "META-INF/MANIFEST.MF")
+ return zipContains(raw, []byte("META-INF/MANIFEST.MF"))
 }
 
 // zipTokenizer holds the source zip file and scanned index.
@@ -54,7 +53,7 @@ type zipTokenizer struct {
 
 // next returns the next file name from the zip headers.
 // https://web.archive.org/web/20191129114319/https://users.cs.jmu.edu/buchhofp/forensics/formats/pkzip.html
-func (t *zipTokenizer) next() (fileName string) {
+func (t *zipTokenizer) next() (fileName []byte) {
  if t.i > len(t.in) {
  return
  }
@@ -74,15 +73,15 @@ func (t *zipTokenizer) next() (fileName string) {
  return
  }
  t.i += fNameOffset + fNameLen
- return string(in[fNameOffset : fNameOffset+fNameLen])
+ return in[fNameOffset : fNameOffset+fNameLen]
 }
 
 // zipContains returns true if the zip file headers from in contain any of the paths.
-func zipContains(in []byte, paths ...string) bool {
+func zipContains(in []byte, paths ...[]byte) bool {
  t := zipTokenizer{in: in}
- for i, tok := 0, t.next(); tok != ""; i, tok = i+1, t.next() {
+ for tok := t.next(); len(tok) != 0; tok = t.next() {
  for p := range paths {
- if strings.HasPrefix(tok, paths[p]) {
+ if bytes.HasPrefix(tok, paths[p]) {
  return true
  }
  }