there is a setting on when and wheter to auto delete partial uploads on s3 for multipart.

define the behaviour and exception on when the client uploads data but the old segments are deleted/cleaned by s3

uploading a 30mb lz4 file fails:
object.Exception occurred at ../../../vibe.d-master/source/vibe/core/drivers/libevent2_tcp.d:431 : Connection error while writing to TCPConnection.

create a range based interface to read/write data in a streaming way.

Compiling examples produces:

source/app.d(105,18): Error: function vibe.aws.s3.S3.upload (string resource, RandomAccessStream input, string contentType = "application/octet-stream", StorageClass storageClass = "STANDARD", ulong chunkSize = 524288LU) is not callable using argument types (string, FileStream, string)

Although FileStream implements RandomAccessStream

Using the in another program produces the same error.
Any thoughts?

support multi-part upload
http://docs.aws.amazon.com/AmazonS3/latest/API/mpUploadInitiate.html

I got this error

vibe-s3 0.0.6: building configuration "library"...
Enhanced memory security is enabled.
Memory debugger enabled
../../../../.dub/packages/vibe-s3-0.0.6/vibe-s3/source/vibe/aws/aws.d(333,13): Error: none of the overloads of 'maxBufferSize' are callable using argument types (ulong), candidates are:
../../../../.dub/packages/vibe-d-0.7.30/vibe-d/source/vibe/http/common.d(451,19):        vibe.http.common.ChunkedOutputStream.maxBufferSize()
../../../../.dub/packages/vibe-d-0.7.30/vibe-d/source/vibe/http/common.d(453,17):        vibe.http.common.ChunkedOutputStream.maxBufferSize(uint bytes)
../../../../.dub/packages/vibe-s3-0.0.6/vibe-s3/source/vibe/aws/s3.d(269,24): Error: cannot implicitly convert expression (input.leastSize()) of type ulong to uint
../../../../.dub/packages/vibe-s3-0.0.6/vibe-s3/source/vibe/aws/s3.d(280,25): Error: cannot implicitly convert expression (input.leastSize()) of type ulong to uint
dmd failed with exit code 1.

it might be a good idea to fix the aws regions within an enum.
they rarely change and are easy to upgrade.
this would prevent wrong region names

Code at line 160 in s3.d

super(bucket ~ ".s3-" ~ region ~ ".amazonaws.com", region, "s3", credsSource, config);

has hard-coded the s3 endPoint/bucket URL. You could make it configurable in ClientConfiguration such that it can be used for other services that support s3 API. Moreover, there are some local fake s3 servers for development which do not support the hard-coded URL.

I know vibe-s3 was developed for s3 solely but using s3 for development online is mostly unreliable/difficult. There are several servers that emulate s3 API calls for local development so hard coded params does not make it easy to use vibe-s3 without changing code.

There are several implementation such as;

fakes3 - https://github.com/jubos/fake-s3
a Rackspace object storage API which emulates S3

There are also a number of enterprise solutions listed here with S3 API support http://www.s3-client.com/s3-compatible-storage-solutions.html

It seems such functionality is already supported in code. How about a demo on them?

support a way of setting life cycle settings
http://docs.aws.amazon.com/AmazonS3/latest/dev/mpuoverview.html
http://docs.aws.amazon.com/AmazonS3/latest/dev/object-lifecycle-mgmt.html

I was just wondering.. do you really depend on features only present in vibe-d ver. 0.8.0-beta.5? Because of this high version it is not possible to use this dub configuration in projects that use older and more stable versions...

requests has a nicer API and also supports vibe.
it starts to become the de-facto standard for doing http requests in D.