shares: use more efficient make([]byte, N) not bytes.Repeat([]byte{0}…

…, N)

This change uses Go's memory model guarantees that the zero value memory
shall be zeroed and hence to create a slice of 0 bytes aka padding,
we can simply use make([]byte, N) which is much more efficient than
bytes.Repeat([]byte{0}, N) and the benchmarks show this improvement:

```shell
$ benchstat before.txt after.txt
name               old time/op    new time/op    delta
PaddingVsRepeat-8     674ns ± 1%     538ns ± 0%  -20.22%  (p=0.000 n=9+9)

name               old alloc/op   new alloc/op   delta
PaddingVsRepeat-8    1.31kB ± 0%    0.83kB ± 0%  -36.59%  (p=0.000 n=10+10)

name               old allocs/op  new allocs/op  delta
PaddingVsRepeat-8      12.0 ± 0%      11.0 ± 0%   -8.33%  (p=0.000 n=10+10)
```

shares/padding.go

@@ -1,7 +1,6 @@
 package shares
 
 import (
-	"bytes"
 	"errors"
 
 	"github.com/celestiaorg/go-square/namespace"
@@ -20,7 +19,7 @@ func NamespacePaddingShare(ns namespace.Namespace, shareVersion uint8) (Share, e
 	if err := b.WriteSequenceLen(0); err != nil {
 		return Share{}, err
 	}
-	padding := bytes.Repeat([]byte{0}, FirstSparseShareContentSize)
+	padding := make([]byte, FirstSparseShareContentSize)
 	b.AddData(padding)
 
 	share, err := b.Build()

shares/utils.go

@@ -31,8 +31,7 @@ func zeroPadIfNecessary(share []byte, width int) (padded []byte, bytesOfPadding
 	}
 
 	missingBytes := width - oldLen
-	padByte := []byte{0}
-	padding := bytes.Repeat(padByte, missingBytes)
+	padding := make([]byte, missingBytes)
 	share = append(share, padding...)
 	return share, missingBytes
 }