mirror of
https://github.com/gohugoio/hugo.git
synced 2024-11-14 20:37:55 -05:00
e67886c038
And remove now unsued hashing funcs.
299 lines
6.5 KiB
Go
299 lines
6.5 KiB
Go
// Copyright 2019 The Hugo Authors. All rights reserved.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
|
|
package helpers
|
|
|
|
import (
|
|
"bytes"
|
|
"fmt"
|
|
"io"
|
|
"net"
|
|
"os"
|
|
"path/filepath"
|
|
"sort"
|
|
"strings"
|
|
"unicode"
|
|
"unicode/utf8"
|
|
|
|
bp "github.com/gohugoio/hugo/bufferpool"
|
|
|
|
"github.com/spf13/afero"
|
|
|
|
"github.com/jdkato/prose/transform"
|
|
)
|
|
|
|
// FilePathSeparator as defined by os.Separator.
|
|
const FilePathSeparator = string(filepath.Separator)
|
|
|
|
// TCPListen starts listening on a valid TCP port.
|
|
func TCPListen() (net.Listener, *net.TCPAddr, error) {
|
|
l, err := net.Listen("tcp", ":0")
|
|
if err != nil {
|
|
return nil, nil, err
|
|
}
|
|
addr := l.Addr()
|
|
if a, ok := addr.(*net.TCPAddr); ok {
|
|
return l, a, nil
|
|
}
|
|
l.Close()
|
|
return nil, nil, fmt.Errorf("unable to obtain a valid tcp port: %v", addr)
|
|
}
|
|
|
|
// FirstUpper returns a string with the first character as upper case.
|
|
func FirstUpper(s string) string {
|
|
if s == "" {
|
|
return ""
|
|
}
|
|
r, n := utf8.DecodeRuneInString(s)
|
|
return string(unicode.ToUpper(r)) + s[n:]
|
|
}
|
|
|
|
// UniqueStrings returns a new slice with any duplicates removed.
|
|
func UniqueStrings(s []string) []string {
|
|
unique := make([]string, 0, len(s))
|
|
for i, val := range s {
|
|
var seen bool
|
|
for j := 0; j < i; j++ {
|
|
if s[j] == val {
|
|
seen = true
|
|
break
|
|
}
|
|
}
|
|
if !seen {
|
|
unique = append(unique, val)
|
|
}
|
|
}
|
|
return unique
|
|
}
|
|
|
|
// UniqueStringsReuse returns a slice with any duplicates removed.
|
|
// It will modify the input slice.
|
|
func UniqueStringsReuse(s []string) []string {
|
|
result := s[:0]
|
|
for i, val := range s {
|
|
var seen bool
|
|
|
|
for j := 0; j < i; j++ {
|
|
if s[j] == val {
|
|
seen = true
|
|
break
|
|
}
|
|
}
|
|
|
|
if !seen {
|
|
result = append(result, val)
|
|
}
|
|
}
|
|
return result
|
|
}
|
|
|
|
// UniqueStringsSorted returns a sorted slice with any duplicates removed.
|
|
// It will modify the input slice.
|
|
func UniqueStringsSorted(s []string) []string {
|
|
if len(s) == 0 {
|
|
return nil
|
|
}
|
|
ss := sort.StringSlice(s)
|
|
ss.Sort()
|
|
i := 0
|
|
for j := 1; j < len(s); j++ {
|
|
if !ss.Less(i, j) {
|
|
continue
|
|
}
|
|
i++
|
|
s[i] = s[j]
|
|
}
|
|
|
|
return s[:i+1]
|
|
}
|
|
|
|
// ReaderToBytes takes an io.Reader argument, reads from it
|
|
// and returns bytes.
|
|
func ReaderToBytes(lines io.Reader) []byte {
|
|
if lines == nil {
|
|
return []byte{}
|
|
}
|
|
b := bp.GetBuffer()
|
|
defer bp.PutBuffer(b)
|
|
|
|
b.ReadFrom(lines)
|
|
|
|
bc := make([]byte, b.Len())
|
|
copy(bc, b.Bytes())
|
|
return bc
|
|
}
|
|
|
|
// ReaderToString is the same as ReaderToBytes, but returns a string.
|
|
func ReaderToString(lines io.Reader) string {
|
|
if lines == nil {
|
|
return ""
|
|
}
|
|
b := bp.GetBuffer()
|
|
defer bp.PutBuffer(b)
|
|
b.ReadFrom(lines)
|
|
return b.String()
|
|
}
|
|
|
|
// ReaderContains reports whether subslice is within r.
|
|
func ReaderContains(r io.Reader, subslice []byte) bool {
|
|
if r == nil || len(subslice) == 0 {
|
|
return false
|
|
}
|
|
|
|
bufflen := len(subslice) * 4
|
|
halflen := bufflen / 2
|
|
buff := make([]byte, bufflen)
|
|
var err error
|
|
var n, i int
|
|
|
|
for {
|
|
i++
|
|
if i == 1 {
|
|
n, err = io.ReadAtLeast(r, buff[:halflen], halflen)
|
|
} else {
|
|
if i != 2 {
|
|
// shift left to catch overlapping matches
|
|
copy(buff[:], buff[halflen:])
|
|
}
|
|
n, err = io.ReadAtLeast(r, buff[halflen:], halflen)
|
|
}
|
|
|
|
if n > 0 && bytes.Contains(buff, subslice) {
|
|
return true
|
|
}
|
|
|
|
if err != nil {
|
|
break
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// GetTitleFunc returns a func that can be used to transform a string to
|
|
// title case.
|
|
//
|
|
// # The supported styles are
|
|
//
|
|
// - "Go" (strings.Title)
|
|
// - "AP" (see https://www.apstylebook.com/)
|
|
// - "Chicago" (see https://www.chicagomanualofstyle.org/home.html)
|
|
// - "FirstUpper" (only the first character is upper case)
|
|
// - "None" (no transformation)
|
|
//
|
|
// If an unknown or empty style is provided, AP style is what you get.
|
|
func GetTitleFunc(style string) func(s string) string {
|
|
switch strings.ToLower(style) {
|
|
case "go":
|
|
//lint:ignore SA1019 keep for now.
|
|
return strings.Title
|
|
case "chicago":
|
|
tc := transform.NewTitleConverter(transform.ChicagoStyle)
|
|
return tc.Title
|
|
case "none":
|
|
return func(s string) string { return s }
|
|
case "firstupper":
|
|
return FirstUpper
|
|
default:
|
|
tc := transform.NewTitleConverter(transform.APStyle)
|
|
return tc.Title
|
|
}
|
|
}
|
|
|
|
// HasStringsPrefix tests whether the string slice s begins with prefix slice s.
|
|
func HasStringsPrefix(s, prefix []string) bool {
|
|
return len(s) >= len(prefix) && compareStringSlices(s[0:len(prefix)], prefix)
|
|
}
|
|
|
|
// HasStringsSuffix tests whether the string slice s ends with suffix slice s.
|
|
func HasStringsSuffix(s, suffix []string) bool {
|
|
return len(s) >= len(suffix) && compareStringSlices(s[len(s)-len(suffix):], suffix)
|
|
}
|
|
|
|
func compareStringSlices(a, b []string) bool {
|
|
if a == nil && b == nil {
|
|
return true
|
|
}
|
|
|
|
if a == nil || b == nil {
|
|
return false
|
|
}
|
|
|
|
if len(a) != len(b) {
|
|
return false
|
|
}
|
|
|
|
for i := range a {
|
|
if a[i] != b[i] {
|
|
return false
|
|
}
|
|
}
|
|
|
|
return true
|
|
}
|
|
|
|
// SliceToLower goes through the source slice and lowers all values.
|
|
func SliceToLower(s []string) []string {
|
|
if s == nil {
|
|
return nil
|
|
}
|
|
|
|
l := make([]string, len(s))
|
|
for i, v := range s {
|
|
l[i] = strings.ToLower(v)
|
|
}
|
|
|
|
return l
|
|
}
|
|
|
|
// IsWhitespace determines if the given rune is whitespace.
|
|
func IsWhitespace(r rune) bool {
|
|
return r == ' ' || r == '\t' || r == '\n' || r == '\r'
|
|
}
|
|
|
|
// PrintFs prints the given filesystem to the given writer starting from the given path.
|
|
// This is useful for debugging.
|
|
func PrintFs(fs afero.Fs, path string, w io.Writer) {
|
|
if fs == nil {
|
|
return
|
|
}
|
|
|
|
afero.Walk(fs, path, func(path string, info os.FileInfo, err error) error {
|
|
if err != nil {
|
|
panic(fmt.Sprintf("error: path %q: %s", path, err))
|
|
}
|
|
path = filepath.ToSlash(path)
|
|
if path == "" {
|
|
path = "."
|
|
}
|
|
fmt.Fprintln(w, path, info.IsDir())
|
|
return nil
|
|
})
|
|
}
|
|
|
|
// FormatByteCount pretty formats b.
|
|
func FormatByteCount(bc uint64) string {
|
|
const (
|
|
Gigabyte = 1 << 30
|
|
Megabyte = 1 << 20
|
|
Kilobyte = 1 << 10
|
|
)
|
|
switch {
|
|
case bc > Gigabyte || -bc > Gigabyte:
|
|
return fmt.Sprintf("%.2f GB", float64(bc)/Gigabyte)
|
|
case bc > Megabyte || -bc > Megabyte:
|
|
return fmt.Sprintf("%.2f MB", float64(bc)/Megabyte)
|
|
case bc > Kilobyte || -bc > Kilobyte:
|
|
return fmt.Sprintf("%.2f KB", float64(bc)/Kilobyte)
|
|
}
|
|
return fmt.Sprintf("%d B", bc)
|
|
}
|