Changelog

Fix url encoding issue for multipart uploads (#139 - thanks @jomatt)
Add downloading a folder example to gcs_get_object() #144
Improve message feedback for downloading objects
gcs_upload() will now retry with a warning if using legacy ACL settings for a bucket with uniform bucket-level access enabled - will retry with predefinedAcl='bucketLevel'
Add gcs_parse_rds() to help parsing RDS objects for gcs_get_object() #146

Let users auth via an email again via gcs_auth
correct file.copy behaviour in gcs_load_all() (#112) - thanks @jasonmhoule
Allow GCS uploads if ACL has been set at bucket level (#111)
Let users set the boundary between simple and resumable upload limit via new option option(googleCloudStorageR.upload_limit) or gcs_upload_set_limit() - default is 5000000L or 5MB (#120)
Suppress warning for when resumable uploads are requesting the upload URL (#120)
Fix ability to upload meta data with objects (#108)
If file has no filesize will message without an error (#123)
If project has no buckets will give NULL and not an error (#100)
Fix gcs_save_all() to also accept bucket level ACL (#129)
Fix listing buckets for projects gcs_list_buckets() (#100)
Add gcs_setup() to help first time setup

Major changes

Update Bucket storageClass to include MULTI_REGIONAL, REGIONAL, and COLDLINE
Fix bug where gcs_load wouldn’t work if file name not “.RData”
Add gcs_first and gcs_last to autosave your file workspace to GCS
Better message feedback on file sizes
Add gcs_save_all and gcs_load_all which will zip, save/load and upload/download a directory
Add use of _gcssave.yaml file to control gcs_first/last behaviour
Allow passing a bucket object to functions that expect a bucket name (#76)
remove now unsupported travis environment argument
Add support for subscribing bucket changes to Google Pub/Sub

Correct metadata upload (#55 - thanks AndrewMarritt)
Let gcs_object_metaname not require name so it can be reused (#56 - thanks seandavi)
Add support for gs:// style URLs for object names (#57 - thanks seandavi)
Add what the public URL would be to an objects print method (#59 - thanks mwhitaker)
Add check to gcs_get_bucket() to only expect length 1 character vectors for bucket name. (#60)
Add paging for gcs_object_list (#58 - thanks @G3rtjan)
Add saveToDisk option to gcs_load (#52 - thanks @tomsing1)
Supply your own parse function to gcs_get_object() (#63 - thanks @nkeriks)
Support for prefix and delimiter in gcs_object_list to filter objects listed (#68)
Default permissions on new buckets lets you write to them afterwards `projectPrivate (#62)
gcs_get_object now supports downloads over 2GB (#69)
Add support for signed URLs (#54 - thanks seandavi)

Fix bug where you can’t rename objects via name in gcs_upload
Add gcs_save to store R session data in cloud
Add gcs_load to restore session data stored with gcs_save
Fix resetting of options(googleAuthR.rawResponse = TRUE) when using gcs_get_object
Add URLencoding to object_name in gcs_get_object etc.
gcs_global_bucket and gcs_get_global_bucket to set global bucket name
Add options to set environment variables for API keys, default buckets etc.
Add print S3 methods for objects, buckets and ACL
Add option to use your own write function for R object file uploads in gcs_upload
Add option to include object metadata in uploads via gcs_metadata_object
Add resumable uploads to gcs_upload to allow uploads over 5MB, limit now 5TB
Add retry uploads via gcs_retry_upload
Add delete object via gcs_delete_object
Add gcs_source to source .R files from GCS
Add versioning and lifecycle management to create and update bucket functions

Initial release