Add DigestInputStream utils #110
Conversation
There was a problem hiding this comment.
The API changes look fine to me. As for the performance, it looks to be nearly identical with the one exception of the buffer size. 1024 seems like too conservative of a value to me. I think you'll see a decent speed increase by picking a value like 4096 or 8192 as it aligns better with block sizes of most systems. For example, BufferedInputStream.DEFAULT_BUFFER_SIZE is 8192. But if you'd rather tweak that value later after a benchmark is added, that's fine too.
| @@ -305,7 +302,6 @@ class FileSpec extends FlatSpec with BeforeAndAfterEach with Matchers { | |||
| } | |||
|
|
|||
| it should "compute correct checksum for non-zero length string" in { | |||
| implicit val charset = StandardCharsets.UTF_8 | |||
There was a problem hiding this comment.
I added this because the default for writeText is Charset.defaultCharset(), which is system dependent. Otherwise, technically this test can fail on somebody else's machine if their default charset is something unusual.
|
Unfortunately, I wasn't doing anything too sophisticated. Basically using really large files (also numerous files) to ensure that all the processing time was indeed being spent in I was using already existing files on my system to test with, but I suppose we could just generate large files programmatically and test with that. I haven't used JMH yet, but I'd like to get familiar with it. If you're not going to be working on #62 in the near future, I could take a look at it if you want. I wouldn't be able to do it right away, but maybe sometime this month. |
|
@solicode : I don't have any plans for #62 anytime soon. So please feel free to contribute. Look at the benchmarks sub-project: https://github.com/pathikrit/better-files/tree/master/benchmarks/src/test/scala/better/files |
No description provided.