The idea is to see whether it is worthwhile to implement special cases.
On my Mac Book Pro:
BenchmarkEncode 1000 1689656 ns/op 36.69 MB/s
R=nigeltao, minux.ma
CC=golang-dev
https://golang.org/cl/6278050
The basic functionality works. Features to add in future CLs:
- compression
- fast paths for image formats that can be directly expressed in TIFF,
such as RGBA, NRGBA and maybe Gray and Paletted.
R=nigeltao
CC=golang-dev
https://golang.org/cl/5694051
In case the image is read via a tiff.buffer, avoid copying the
data strip before decoding it. Remove corresponding TODO.
Speeds up reading uncompressed images (which is the common case)
and uses much less memory.
benchmark old ns/op new ns/op delta
BenchmarkDecodeCompressed 4619438 4630774 +0.25%
BenchmarkDecodeUncompressed 260809 219875 -15.70%
R=nigeltao
CC=golang-dev
https://golang.org/cl/5683050
Manual edits to README.
Moved from main Go repository, deleted Makefiles.
Tested with go test code.google.com/p/go.image/...
Fixesgolang/go#2797.
R=rsc, rsc, r
CC=golang-dev
https://golang.org/cl/5593053
