Add riscv64 test

.woodpecker.yml

@@ -0,0 +1,13 @@
+matrix:
+  platform:
+    - linux/amd64
+    - linux/arm64
+    - linux/riscv64
+
+steps:
+  test:
+    image: gitea.elara.ws/elara6331/golang:latest
+    commands:
+      - go test
+    when:
+      - event: push

README.md

@@ -1,6 +1,7 @@
 # pcre
 
-[![Go Reference](https://pkg.go.dev/badge/go.arsenm.dev/pcre.svg)](https://pkg.go.dev/go.arsenm.dev/pcre)
+[![Go Reference](https://pkg.go.dev/badge/go.elara.ws/pcre.svg)](https://pkg.go.dev/go.elara.ws/pcre)
+[![status-badge](https://ci.elara.ws/api/badges/49/status.svg)](https://ci.elara.ws/49)
 
 This package provides a CGo-free port of the PCRE2 regular expression library. The [lib](lib) directory contains source code automatically translated from PCRE2's C source. This package wraps that code and provides an interface as close as possible to Go's stdlib [regexp](https://pkg.go.dev/regexp) package
 
@@ -50,4 +51,4 @@ CC=/usr/bin/gcc ccgo -o pcre2_<os>_<arch>.go -pkgname lib -trace-translation-uni
 
 - If cross-compiling, set the `CCGO_CC` variable to to path of the cross-compiler, and the `CCGO_AR` variable to the path of the cross-compiler's `ar` binary. Also, set `TARGET_GOARCH` to the GOARCH you're targeting and `TARGET_GOOS` to the OS you're targeting.
 
-- Once the command completes, two go files will be created. One will start with `pcre2`, the other with `capi`. Copy both of these to the `lib` directory in this repo.
+- Once the command completes, two go files will be created. One will start with `pcre2`, the other with `capi`. Copy both of these to the `lib` directory in this repo.

glob.go

@@ -108,7 +108,7 @@ func Glob(glob string) ([]string, error) {
 
 	// Join splitDir and add filepath separator. This is the directory that will be searched.
 	dir := filepath.Join(splitDir...)
-	
+
 	if filepath.IsAbs(glob) {
 		dir = string(filepath.Separator) + dir
 	}

glob_test.go

@@ -39,17 +39,17 @@ func TestCompileGlob(t *testing.T) {
 }
 
 func TestGlob(t *testing.T) {
-	err := os.MkdirAll("pcretest/dir1", 0755)
+	err := os.MkdirAll("pcretest/dir1", 0o755)
 	if err != nil {
 		t.Fatal(err)
 	}
 
-	err = os.MkdirAll("pcretest/dir2", 0755)
+	err = os.MkdirAll("pcretest/dir2", 0o755)
 	if err != nil {
 		t.Fatal(err)
 	}
 
-	err = os.MkdirAll("pcretest/test1/dir4", 0755)
+	err = os.MkdirAll("pcretest/test1/dir4", 0o755)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -58,7 +58,7 @@ func TestGlob(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	
+
 	err = touch("pcretest/file2")
 	if err != nil {
 		t.Fatal(err)
@@ -113,9 +113,9 @@ func TestGlob(t *testing.T) {
 }
 
 func touch(path string) error {
-	fl, err := os.OpenFile(path, os.O_CREATE, 0644)
+	fl, err := os.OpenFile(path, os.O_CREATE, 0o644)
 	if err != nil {
 		return err
 	}
 	return fl.Close()
-}
+}

pcre.go

@@ -8,6 +8,7 @@
 package pcre
 
 import (
+	"math"
 	"os"
 	"runtime"
 	"strconv"
@@ -19,6 +20,8 @@ import (
 	"modernc.org/libc"
 )
 
+const Unset = math.MaxUint
+
 // Version returns the version of pcre2 embedded in this library.
 func Version() string { return lib.DPACKAGE_VERSION }
 
@@ -208,7 +211,11 @@ func (r *Regexp) FindSubmatch(b []byte) [][]byte {
 
 	out := make([][]byte, 0, len(match)/2)
 	for i := 0; i < len(match); i += 2 {
-		out = append(out, b[match[i]:match[i+1]])
+		if match[i] == Unset {
+			out = append(out, nil)
+		} else {
+			out = append(out, b[match[i]:match[i+1]])
+		}
 	}
 	return out
 }
@@ -253,7 +260,11 @@ func (r *Regexp) FindAllSubmatch(b []byte, n int) [][][]byte {
 		outMatch := make([][]byte, 0, len(match)/2)
 
 		for i := 0; i < len(match); i += 2 {
-			outMatch = append(outMatch, b[match[i]:match[i+1]])
+			if match[i] == Unset {
+				outMatch = append(outMatch, nil)
+			} else {
+				outMatch = append(outMatch, b[match[i]:match[i+1]])
+			}
 		}
 
 		out[index] = outMatch
@@ -565,67 +576,6 @@ func (r *Regexp) SubexpIndex(name string) int {
 	return int(ret)
 }
 
-type CalloutFlags uint32
-
-const (
-	CalloutStartMatch = CalloutFlags(lib.DPCRE2_CALLOUT_STARTMATCH)
-	CalloutBacktrack  = CalloutFlags(lib.DPCRE2_CALLOUT_BACKTRACK)
-)
-
-type CalloutBlock struct {
-	// Version contains the version number of the block format.
-	// The current version is 2.
-	Version uint32
-
-	// CalloutNumber contains the number of the callout, in the range 0-255.
-	// This is the number that follows "?C". For callouts with string arguments,
-	// this will always be zero.
-	CalloutNumber uint32
-
-	// CaptureTop contains the number of the highest numbered substring
-	// captured so far plus one. If no substrings have yet been captured,
-	// CaptureTop will be set to 1.
-	CaptureTop uint32
-
-	// CaptureLast contains the number of the last substring that was captured.
-	CaptureLast uint32
-
-	// Substrings contains all of the substrings captured so far.
-	Substrings []string
-
-	Mark string
-
-	// Subject contains the string passed to the match function.
-	Subject string
-
-	// StartMatch contains the offset within the subject at which the current match attempt started.
-	StartMatch uint
-
-	// CurrentPosition contains the offset of the current match pointer within the subject.
-	CurrentPosition uint
-
-	// PatternPosition contains the offset within the pattern string to the next item to be matched.
-	PatternPosition uint
-
-	// NextItemLength contains the length of the next item to be processed in the pattern string.
-	NextItemLength uint
-
-	// CalloutStringOffset contains the code unit offset to the start of the callout argument string within the original pattern string.
-	CalloutStringOffset uint
-
-	// CalloutString is the string for the callout. For numerical callouts, this will always be empty.
-	CalloutString string
-
-	// CalloutFlags contains the following flags:
-	// 	CalloutStartMatch
-	// This is set for the first callout after the start of matching for each new starting position in the subject.
-	// 	CalloutBacktrack
-	// This is set if there has been a matching backtrack since the previous callout, or since the start of matching if this is the first callout from a pcre2_match() run.
-	//
-	// Both bits are set when a backtrack has caused a "bumpalong" to a new starting position in the subject. Output
-	CalloutFlags CalloutFlags
-}
-
 // SetCallout sets a callout function that will be called at specified points in the matching operation.
 // fn should return zero if it ran successfully or a non-zero integer to force an error.
 // See https://www.pcre.org/current/doc/html/pcre2callout.html for more information.
@@ -653,12 +603,15 @@ func (r *Regexp) SetCallout(fn func(cb *CalloutBlock) int32) error {
 		calloutStrBytes := unsafe.Slice((*byte)(unsafe.Pointer(ccb.Fcallout_string)), ccb.Fcallout_string_length)
 		cb.CalloutString = string(calloutStrBytes)
 
-		ovecSlice := unsafe.Slice((*lib.Tsize_t)(unsafe.Pointer(ccb.Foffset_vector)), (ccb.Fcapture_top*2)-1)[2:]
-		for i := 0; i < len(ovecSlice); i += 2 {
-			if i+1 >= len(ovecSlice) {
-				cb.Substrings = append(cb.Substrings, cb.Subject[ovecSlice[i]:])
-			} else {
-				cb.Substrings = append(cb.Substrings, cb.Subject[ovecSlice[i]:ovecSlice[i+1]])
+		ovecSlice := unsafe.Slice((*lib.Tsize_t)(unsafe.Pointer(ccb.Foffset_vector)), (ccb.Fcapture_top*2)-1)
+		if len(ovecSlice) > 2 {
+			ovecSlice = ovecSlice[2:]
+			for i := 0; i < len(ovecSlice); i += 2 {
+				if i+1 >= len(ovecSlice) {
+					cb.Substrings = append(cb.Substrings, cb.Subject[ovecSlice[i]:])
+				} else {
+					cb.Substrings = append(cb.Substrings, cb.Subject[ovecSlice[i]:ovecSlice[i+1]])
+				}
 			}
 		}
 

pcre_test.go

@@ -4,6 +4,7 @@ import (
 	"strings"
 	"sync"
 	"testing"
+	"reflect"
 
 	"go.elara.ws/pcre"
 )
@@ -289,3 +290,11 @@ func TestCallout(t *testing.T) {
 		t.Error("expected regular expression to match the string")
 	}
 }
+
+func TestVarnish(t *testing.T) {
+	regex := pcre.MustCompile(`varnish(?: \(Varnish\/([\d.]{1,250})\))?`)
+	matches := regex.FindStringSubmatch("1.1 varnish")
+	if !reflect.DeepEqual(matches, []string{"varnish", ""}) {
+		t.Errorf(`Expected ["varnish" ""], got %q`, matches)
+	}
+}

types.go

@@ -6,33 +6,95 @@ type CompileOption uint32
 
 // Compile option bits
 const (
-	Anchored           = CompileOption(lib.DPCRE2_ANCHORED)
-	AllowEmptyClass    = CompileOption(lib.DPCRE2_ALLOW_EMPTY_CLASS)
-	AltBsux            = CompileOption(lib.DPCRE2_ALT_BSUX)
-	AltCircumflex      = CompileOption(lib.DPCRE2_ALT_CIRCUMFLEX)
-	AltVerbnames       = CompileOption(lib.DPCRE2_ALT_VERBNAMES)
-	AutoCallout        = CompileOption(lib.DPCRE2_AUTO_CALLOUT)
-	Caseless           = CompileOption(lib.DPCRE2_CASELESS)
-	DollarEndOnly      = CompileOption(lib.DPCRE2_DOLLAR_ENDONLY)
-	DotAll             = CompileOption(lib.DPCRE2_DOTALL)
-	DupNames           = CompileOption(lib.DPCRE2_DUPNAMES)
-	EndAnchored        = CompileOption(lib.DPCRE2_ENDANCHORED)
-	Extended           = CompileOption(lib.DPCRE2_EXTENDED)
-	FirstLine          = CompileOption(lib.DPCRE2_FIRSTLINE)
-	Literal            = CompileOption(lib.DPCRE2_LITERAL)
-	MatchInvalidUTF    = CompileOption(lib.DPCRE2_MATCH_INVALID_UTF)
-	MactchUnsetBackref = CompileOption(lib.DPCRE2_MATCH_UNSET_BACKREF)
-	Multiline          = CompileOption(lib.DPCRE2_MULTILINE)
-	NeverBackslashC    = CompileOption(lib.DPCRE2_NEVER_BACKSLASH_C)
-	NeverUCP           = CompileOption(lib.DPCRE2_NEVER_UCP)
-	NeverUTF           = CompileOption(lib.DPCRE2_NEVER_UTF)
-	NoAutoCapture      = CompileOption(lib.DPCRE2_NO_AUTO_CAPTURE)
-	NoAutoPossess      = CompileOption(lib.DPCRE2_NO_AUTO_POSSESS)
-	NoDotStarAnchor    = CompileOption(lib.DPCRE2_NO_DOTSTAR_ANCHOR)
-	NoStartOptimize    = CompileOption(lib.DPCRE2_NO_START_OPTIMIZE)
-	NoUTFCheck         = CompileOption(lib.DPCRE2_NO_UTF_CHECK)
-	UCP                = CompileOption(lib.DPCRE2_UCP)
-	Ungreedy           = CompileOption(lib.DPCRE2_UNGREEDY)
-	UseOffsetLimit     = CompileOption(lib.DPCRE2_USE_OFFSET_LIMIT)
-	UTF                = CompileOption(lib.DPCRE2_UTF)
+	Anchored          = CompileOption(lib.DPCRE2_ANCHORED)
+	AllowEmptyClass   = CompileOption(lib.DPCRE2_ALLOW_EMPTY_CLASS)
+	AltBsux           = CompileOption(lib.DPCRE2_ALT_BSUX)
+	AltCircumflex     = CompileOption(lib.DPCRE2_ALT_CIRCUMFLEX)
+	AltVerbnames      = CompileOption(lib.DPCRE2_ALT_VERBNAMES)
+	AutoCallout       = CompileOption(lib.DPCRE2_AUTO_CALLOUT)
+	Caseless          = CompileOption(lib.DPCRE2_CASELESS)
+	DollarEndOnly     = CompileOption(lib.DPCRE2_DOLLAR_ENDONLY)
+	DotAll            = CompileOption(lib.DPCRE2_DOTALL)
+	DupNames          = CompileOption(lib.DPCRE2_DUPNAMES)
+	EndAnchored       = CompileOption(lib.DPCRE2_ENDANCHORED)
+	Extended          = CompileOption(lib.DPCRE2_EXTENDED)
+	FirstLine         = CompileOption(lib.DPCRE2_FIRSTLINE)
+	Literal           = CompileOption(lib.DPCRE2_LITERAL)
+	MatchInvalidUTF   = CompileOption(lib.DPCRE2_MATCH_INVALID_UTF)
+	MatchUnsetBackref = CompileOption(lib.DPCRE2_MATCH_UNSET_BACKREF)
+	Multiline         = CompileOption(lib.DPCRE2_MULTILINE)
+	NeverBackslashC   = CompileOption(lib.DPCRE2_NEVER_BACKSLASH_C)
+	NeverUCP          = CompileOption(lib.DPCRE2_NEVER_UCP)
+	NeverUTF          = CompileOption(lib.DPCRE2_NEVER_UTF)
+	NoAutoCapture     = CompileOption(lib.DPCRE2_NO_AUTO_CAPTURE)
+	NoAutoPossess     = CompileOption(lib.DPCRE2_NO_AUTO_POSSESS)
+	NoDotStarAnchor   = CompileOption(lib.DPCRE2_NO_DOTSTAR_ANCHOR)
+	NoStartOptimize   = CompileOption(lib.DPCRE2_NO_START_OPTIMIZE)
+	NoUTFCheck        = CompileOption(lib.DPCRE2_NO_UTF_CHECK)
+	UCP               = CompileOption(lib.DPCRE2_UCP)
+	Ungreedy          = CompileOption(lib.DPCRE2_UNGREEDY)
+	UseOffsetLimit    = CompileOption(lib.DPCRE2_USE_OFFSET_LIMIT)
+	UTF               = CompileOption(lib.DPCRE2_UTF)
 )
+
+type CalloutFlags uint32
+
+const (
+	CalloutStartMatch = CalloutFlags(lib.DPCRE2_CALLOUT_STARTMATCH)
+	CalloutBacktrack  = CalloutFlags(lib.DPCRE2_CALLOUT_BACKTRACK)
+)
+
+// CalloutBlock contains the data passed to callout functions
+type CalloutBlock struct {
+	// Version contains the version number of the block format.
+	// The current version is 2.
+	Version uint32
+
+	// CalloutNumber contains the number of the callout, in the range 0-255.
+	// This is the number that follows "?C". For callouts with string arguments,
+	// this will always be zero.
+	CalloutNumber uint32
+
+	// CaptureTop contains the number of the highest numbered substring
+	// captured so far plus one. If no substrings have yet been captured,
+	// CaptureTop will be set to 1.
+	CaptureTop uint32
+
+	// CaptureLast contains the number of the last substring that was captured.
+	CaptureLast uint32
+
+	// Substrings contains all of the substrings captured so far.
+	Substrings []string
+
+	Mark string
+
+	// Subject contains the string passed to the match function.
+	Subject string
+
+	// StartMatch contains the offset within the subject at which the current match attempt started.
+	StartMatch uint
+
+	// CurrentPosition contains the offset of the current match pointer within the subject.
+	CurrentPosition uint
+
+	// PatternPosition contains the offset within the pattern string to the next item to be matched.
+	PatternPosition uint
+
+	// NextItemLength contains the length of the next item to be processed in the pattern string.
+	NextItemLength uint
+
+	// CalloutStringOffset contains the code unit offset to the start of the callout argument string within the original pattern string.
+	CalloutStringOffset uint
+
+	// CalloutString is the string for the callout. For numerical callouts, this will always be empty.
+	CalloutString string
+
+	// CalloutFlags contains the following flags:
+	// 	CalloutStartMatch
+	// This is set for the first callout after the start of matching for each new starting position in the subject.
+	// 	CalloutBacktrack
+	// This is set if there has been a matching backtrack since the previous callout, or since the start of matching if this is the first callout from a pcre2_match() run.
+	//
+	// Both bits are set when a backtrack has caused a "bumpalong" to a new starting position in the subject.
+	CalloutFlags CalloutFlags
+}