Created
July 22, 2020 03:18
-
-
Save rbranson/4c4adbfd4b5b644b35375d8f19ecff50 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
// delimbuf is a byte buffer that tracks the "records" of data | |
// that is appended to it. | |
type delimbuf struct { | |
buf []byte | |
// Record offsets in the buffer are tracked as offsets relative | |
// to base. The purpose of this is to allow Shift() to quickly | |
// remove records without an O(N) operation to rebase the offsets | |
// stored in delims based on the shifted buffer contents. | |
// | |
// Particularly obnoxious readers will notice that there is no | |
// handling of wrap-arounds, which will almost certainly panic. | |
// This would require pushing 9.6EiB through the buffer, which | |
// is a little far-fetched. Perhaps my attitude would be different | |
// if this was avionics and not a pipeline of JSON serdes. | |
delims []int64 | |
base int64 | |
} | |
func (d *delimbuf) unbase(bo int64) int { | |
return int(bo - d.base) | |
} | |
func (d *delimbuf) rebase(offset int) int64 { | |
return int64(offset) + d.base | |
} | |
func (d *delimbuf) recordLen(index int) int { | |
this := d.unbase(d.delims[index]) | |
next := len(d.buf) | |
lastidx := len(d.delims) - 1 | |
if index < lastidx { | |
next = d.unbase(d.delims[index+1]) | |
} | |
return next - this | |
} | |
// Append adds record src to the buffer | |
func (d *delimbuf) Append(src []byte) { | |
if d.buf == nil { | |
d.buf = make([]byte, 0, len(src)) | |
} | |
if d.delims == nil { | |
d.delims = make([]int64, 0, 1) | |
} | |
offset := d.rebase(len(d.buf)) | |
d.delims = append(d.delims, offset) | |
d.buf = append(d.buf, src...) | |
} | |
// BufferLen returns the number of bytes in the buffer | |
func (d *delimbuf) BufferLen() int { | |
return len(d.buf) | |
} | |
// Len returns the number of buffered records | |
func (d *delimbuf) Len() int { | |
return len(d.delims) | |
} | |
// Shift removes count records from front of the buffer | |
func (d *delimbuf) Shift(count int) { | |
if len(d.buf) < count || len(d.delims) < count { | |
panic("shift exceeds available elements") | |
} | |
idx := count - 1 | |
nbytes := d.unbase(d.delims[idx]) + d.recordLen(idx) | |
d.delims = d.delims[count:] | |
if nbytes > 0 { | |
d.base += int64(nbytes) | |
d.buf = d.buf[nbytes:] | |
} | |
} | |
// Get returns a slice of the buffer for record at index | |
func (d *delimbuf) Get(index int) []byte { | |
if d.buf == nil || d.delims == nil { | |
panic("index out of bounds") | |
} | |
from := d.unbase(d.delims[index]) | |
to := from + d.recordLen(index) | |
return d.buf[from:to] | |
} | |
// IndexAtOffset returns the record index for the buffer position | |
// specified by offset. This is an ~O(log n) operation where n is | |
// the number of records in the buffer. | |
func (d *delimbuf) IndexForOffset(offset int) int { | |
if offset >= len(d.buf) { | |
panic("offset is beyond buffer") | |
} | |
// Do a right-most binary search to find the record, taking advantage | |
// of the monotonic nature of delims. | |
left := 0 | |
right := len(d.delims) | |
for left < right { | |
mid := (left + right) / 2 // remainder is truncated | |
if d.unbase(d.delims[mid]) > offset { | |
right = mid | |
} else { | |
left = mid + 1 | |
} | |
} | |
return right - 1 | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment