fmt: allow padding and minus flags at the same time

Existing implementation did not allow setting both padding and
minus flags at the same time because standard formatting does
not allow that. But custom Formatter interface implementations
might have use of it. This change moves the check from the
place flags are parsed to where they are used in standard
formatting.

Fixes golang#61784

src/fmt/fmt_test.go

@@ -1501,6 +1501,7 @@ var flagtests = []struct {
 	{"%-+1.2a", "[%+-1.2a]"},
 	{"%-+1.2abc", "[%+-1.2a]bc"},
 	{"%-1.2abc", "[%-1.2a]bc"},
+	{"%-0abc", "[%-0a]bc"},
 }
 
 func TestFlagParser(t *testing.T) {
@@ -1827,6 +1828,7 @@ var formatterFlagTests = []struct {
 	{"%-+1.2a", flagPrinter{}, "[%+-1.2a]"},
 	{"%-+1.2abc", flagPrinter{}, "[%+-1.2a]bc"},
 	{"%-1.2abc", flagPrinter{}, "[%-1.2a]bc"},
+	{"%-0abc", flagPrinter{}, "[%-0a]bc"},
 
 	// composite values with the 'a' verb
 	{"%a", [1]flagPrinter{}, "[[%a]]"},
@@ -1841,6 +1843,7 @@ var formatterFlagTests = []struct {
 	{"%-+1.2a", [1]flagPrinter{}, "[[%+-1.2a]]"},
 	{"%-+1.2abc", [1]flagPrinter{}, "[[%+-1.2a]]bc"},
 	{"%-1.2abc", [1]flagPrinter{}, "[[%-1.2a]]bc"},
+	{"%-0abc", [1]flagPrinter{}, "[[%-0a]]bc"},
 
 	// simple values with the 'v' verb
 	{"%v", flagPrinter{}, "[%v]"},
@@ -1855,6 +1858,7 @@ var formatterFlagTests = []struct {
 	{"%-+1.2v", flagPrinter{}, "[%+-1.2v]"},
 	{"%-+1.2vbc", flagPrinter{}, "[%+-1.2v]bc"},
 	{"%-1.2vbc", flagPrinter{}, "[%-1.2v]bc"},
+	{"%-0vbc", flagPrinter{}, "[%-0v]bc"},
 
 	// composite values with the 'v' verb.
 	{"%v", [1]flagPrinter{}, "[[%v]]"},
@@ -1869,6 +1873,7 @@ var formatterFlagTests = []struct {
 	{"%-+1.2v", [1]flagPrinter{}, "[[%+-1.2v]]"},
 	{"%-+1.2vbc", [1]flagPrinter{}, "[[%+-1.2v]]bc"},
 	{"%-1.2vbc", [1]flagPrinter{}, "[[%-1.2v]]bc"},
+	{"%-0vbc", [1]flagPrinter{}, "[[%-0v]]bc"},
 }
 
 func TestFormatterFlags(t *testing.T) {

src/fmt/format.go

@@ -77,7 +77,8 @@ func (f *fmt) writePadding(n int) {
 	}
 	// Decide which byte the padding should be filled with.
 	padByte := byte(' ')
-	if f.zero {
+	// Zero padding is allowed only to the left.
+	if f.zero && !f.minus {
 		padByte = byte('0')
 	}
 	// Fill padding with padByte.
@@ -225,7 +226,7 @@ func (f *fmt) fmtInteger(u uint64, base int, isSigned bool, verb rune, digits st
 			f.zero = oldZero
 			return
 		}
-	} else if f.zero && f.widPresent {
+	} else if f.zero && !f.minus && f.widPresent { // Zero padding is allowed only to the left.
 		prec = f.wid
 		if negative || f.plus || f.space {
 			prec-- // leave room for sign
@@ -582,7 +583,8 @@ func (f *fmt) fmtFloat(v float64, size int, verb rune, prec int) {
 	if f.plus || num[0] != '+' {
 		// If we're zero padding to the left we want the sign before the leading zeros.
 		// Achieve this by writing the sign out and then padding the unsigned number.
-		if f.zero && f.widPresent && f.wid > len(num) {
+		// Zero padding is allowed only to the left.
+		if f.zero && !f.minus && f.widPresent && f.wid > len(num) {
 			f.buf.writeByte(num[0])
 			f.writePadding(f.wid - len(num))
 			f.buf.write(num[1:])

src/fmt/print.go

@@ -1048,12 +1048,11 @@ formatLoop:
 			case '#':
 				p.fmt.sharp = true
 			case '0':
-				p.fmt.zero = !p.fmt.minus // Only allow zero padding to the left.
+				p.fmt.zero = true
 			case '+':
 				p.fmt.plus = true
 			case '-':
 				p.fmt.minus = true
-				p.fmt.zero = false // Do not pad with zeros to the right.
 			case ' ':
 				p.fmt.space = true
 			default: