Query string sorting in canonicalisation

[endgame]

@nitinprakash96 please check if this works for you.

This PR fixes the ordering of query parameters during request signing. They should now sort correctly in all instances. Although I've added tests, I really don't want to break anything here. If you use Amazonka with an API that uses query parameters in its request format (potentially: IAM, SNS, SimpleDB, CloudWatch, SES, RDS, STS, Elastic Load Balancing (v1 or v2), SQS, DocumentDB, CloudSearch, Redshift, Application Auto Scaling, CloudFormation, Neptune, Elasticache (control plane), Elastic Beanstalk, or Import/Export), I'd love to hear from you.

Fixes #965

[endgame]

@nitinprakash96 I should also say, S3 Select is not being offered to new customers, so while I no longer get the signing error, I do get this:

[ServiceError] {
  service    = "S3"
  status     = 405 "Method Not Allowed"
  code       = "MethodNotAllowed"
  message    = Just "The specified method is not allowed against this resource."
  request-id = Just "E26Y9BHFKJWPTA3C"
}

I cannot test further, so please let me know if it works for you.

[nitinprakash96]

Thanks for looking into this @endgame . Appreciate it. I can test it out over the weekend if you don't mind.

[endgame]

@nitinprakash96 I'd appreciate that a lot. All I can do is check that signature verification passes, but I can't tell whether S3 Select actually works now.

[nitinprakash96]

While the signature issue is resolved by this error, I run into the following exception:

ghci> selectObjectContent ("nitin-testing") ("airtravel.csv") "select * from S3Object limit 5"
[Client Request] {
  host      = nitin-testing.s3.us-east-1.amazonaws.com:443
  secure    = True
  method    = POST
  target    = Nothing
  timeout   = ResponseTimeoutMicro 70000000
  redirects = 0
  path      = /airtravel.csv
  query     = ?select=&select-type=2
  headers   =  <headers>
  body      = non-printable strict ByteString (480 bytes)
}
[Client Response] {
  status  = 200 OK
  headers = <headers>
}
[SerializeError] {
  service = S3
  status  = 200 OK
  message = Data.Conduit.Text.decode: Error decoding stream of UTF-8 bytes. Error encountered in stream at offset 3. Encountered at byte sequence "\211\NUL\NUL\NUL"
  body    = Just non-printable lazy ByteString (467 bytes)
}
*** Exception: SerializeError (SerializeError' {abbrev = Abbrev {fromAbbrev = "S3"}, status = Status {statusCode = 200, statusMessage = "OK"}, body = Just "\NUL\NUL\NUL\211\NUL\NUL\NULU\176v\DEL\155\r:message-type\a\NUL\ENQevent\v:event-type\a\NUL\aRecords\r:content-type\a\NUL\CANapplication/octet-streamJAN,  340,  360,  417\nFEB,  318,  342,  391\nMAR,  362,  406,  419\nAPR,  348,  396,  461\nMAY,  363,  420,  472\n\226\184<\178\NUL\NUL\NUL\200\NUL\NUL\NULCS\146lY\r:message-type\a\NUL\ENQevent\v:event-type\a\NUL\ENQStats\r:content-type\a\NUL\btext/xml<Stats><BytesScanned>321</BytesScanned><BytesProcessed>321</BytesProcessed><BytesReturned>110</BytesReturned></Stats>\214\238T\191\NUL\NUL\NUL8\NUL\NUL\NUL(\193\198\132\212\r:message-type\a\NUL\ENQevent\v:event-type\a\NUL\ETXEnd\207\151\211\146", message = "Data.Conduit.Text.decode: Error decoding stream of UTF-8 bytes. Error encountered in stream at offset 3. Encountered at byte sequence \"\\211\\NUL\\NUL\\NUL\""})

But I'm not sure if this is due to this PR or the calling code is incorrect. I'm using the same snippet as described in the corresponding issue.
Also, I tried against multiple CSVs and all of them resulted in the same error. Just that the encountered byte sequence was different each time.