> + String key = blob.getMetadata().getName();
> + Date creationDate = null;
> + BlobEntity oldBlobEntity = findBlobById(containerName, key);
> + if (oldBlobEntity != null) {
> + creationDate = oldBlobEntity.getCreationDate();
> + deleteBlob(containerName, key);
> + }
> + BlobEntity blobEntity = blobToBlobEntity.apply(blob);
> +
> blobEntity.getPayload().setChunks(storeData(blob.getPayload().openStream()));
> +
> blobEntity.setContainerEntity(containerRepository.findContainerByName(containerName));
> + blobEntity.setKey(key);
> + blobEntity.setBlobAccess(blobAccess);
> + blobEntity.setCreationDate(creationDate);
> +
> +
> + HashCode hash =
> ByteStreams2.hashAndClose(blob.getPayload().openStream(), Hashing.md5());
Somewhat of a small problem with this.
In a lot of cases, the user has already checked the hash of the stream they are
trying to upload to a cloud/jdbc/etc. Which means that by having us to it
again, performance is reduced overall. That is really the point of returning
the hash.
I am mostly OK with keeping it this way, but maybe explicitly document that the
check happens here too? Either that or remove the integrity check. Removing it
might be better. WDYT?
---
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds-labs/pull/176/files#r32112760
