sgeb / acdcli Goto Github PK

% acdcli get Documents/Subfolder/file.txt

It should:

• download the content of file Documents/Subfolder/file.txt and store as file.txt in the current working directory
• handle prefix / as optional in the command argument
• print an error if file file.txt already exists in current working directory
• print an error in case a suffix / is present
• print an error in case no such file exists on the Amazon Cloud Drive

Copy a single file for now, support globs and whole folders at a later point.

For example:

% acdcli cp /Documents/Samples/sample.jpg /Documents/Pictures

Currently acdcli storage displays used storage, grouped by category. It should also display quota information.

% acdcli get /Documents/Subfolder/file.txt ~/Desktop/anotherfile.txt

It should:

• download the content of file Documents/Subfolder/file.txt and store as anotherfile.txt in directory $HOME/Desktop
• handle prefix / as optional in the remote node path argument
• respect common path rules for target directory (~, /, ., etc)
• print an error if target directory does not exist
• print an error if file anotherfile.txt already exists in target directory
• print an error in case a suffix / is present in remote node path
• print an error in case no such file exists on the Amazon Cloud Drive

Depends on issue #7.

Hi,

I noticed that you plan to embed Amazon's client id and secret in the binary. While this is convinient, this is also insecure.

wmn@cratos ~/code/src/github.com/sgeb/acdcli [master] ± % ACD_API_CLIENTID="amzn1.application-oa2-client.this_is_a_hash" ACD_API_SECRET="this_is_a_secret" make dev
==> Getting dependencies...
>> Getting package github.com/sgeb/go-acd
>> Getting package golang.org/x/oauth2
>> Getting package github.com/mitchellh/cli
>> Getting package github.com/dustin/go-humanize
>> Getting package github.com/mitchellh/gox
package github.com/mitchellh/cli
    imports golang.org/x/crypto/ssh/terminal: golang.org/x/crypto is a custom import path for https://go.googlesource.com/crypto, but /Users/wmn/code/src/golang.org/x/crypto is checked out from https://code.google.com/p/go.crypto
>> Setting github.com/dustin/go-humanize to version
>> Setting golang.org/x/oauth2 to version
>> Setting github.com/mitchellh/cli to version
>> Setting github.com/sgeb/go-acd to version
>> Setting github.com/mitchellh/gox to version
>> Building package github.com/sgeb/go-acd
>> Building package golang.org/x/oauth2
>> Building package github.com/mitchellh/cli
>> Building package github.com/dustin/go-humanize
>> Building package github.com/mitchellh/gox
>> All Done
==> Removing old directory...
==> Building...
Number of parallel builds: 8

-->    darwin/amd64: github.com/sgeb/acdcli/acdcli

==> Results:
total 15024
-rwxr-xr-x  1 wmn  staff   7.3M May 22 10:25 acdcli
wmn@cratos ~/code/src/github.com/sgeb/acdcli [master] ± % strings ~/code/bin/acdcli| grep 'a_hash\|a_secret'
this_is_a_secret
amzn1.application-oa2-client.this_is_a_hash

Although the secret does not have a known guessable prefix, it is always 64 chars which is easy guessable.

Take a look at https://github.com/caseymrm/drivesink, they have the client/secret hosted on a website and the users only get their access token and refresh token.

Move a single file for now, support globs and whole folders at a later point.

For example:

% acdcli mv /Documents/Samples/sample.jpg /Documents/Pictures

• List of files sortable by size
• Display date of last modification
• Cache metadata locally, update periodically using eTag
• Extract logTransport into separate library
• Display periodic updates on download/upload progress, speed and ETA
• Support CLI options to switch between 1000 and 1024 size units (i.e. MB vs MiB). I prefer the former, Amazon Cloud Drive displays the latter. Currently uses the latter to not confuse users where acdcli would display different sizes from amazon on the web.
• ls accepts multiple parameters on the commandline
• create intermediate directories for the target of a download
• Commands share and unshare
• Make sure paths work on windows (see os.IsPathSeparator)
• Command sync

% acdcli get /Documents/Subfolder/file.txt ~/Desktop/

It should:

• download the content of file Documents/Subfolder/file.txt and store as file.txt in directory $HOME/Desktop
• handle prefix / as optional in the remote node path argument
• handle suffix / as optional in the target directory argument
• respect common path rules for target directory (~, /, ., etc)
• print an error if target directory does not exist
• print an error if file file.txt already exists in target directory
• print an error in case a suffix / is present in remote node path
• print an error in case no such file exists on the Amazon Cloud Drive

Depends on issue #6.

Passing the -d/--debug argument enables debug output.

% acdacli info /Documents/Subfolder/file.txt
[... metadata as json ...]

It should:

• display the metadata of Documents/Subfolder/file.txt as pretty-printed JSON
• handle prefix / as optional in the command argument
• print an error in case a suffix / is present and the deepest node is not a directory
• print an error in case Documents/Subfolder/file.txt does not exist

% acdcli ls /Documents/Subfolder
AnotherFolder/
example.jpg
sample1.txt
test.txt

It should:

• list the content of folder Documents/Subfolder if it is indeed a folder
• list the nodes using the same order rules as in issue #2
• list the file if the deepest node is a file
• handle prefix / as optional in the command argument
• print an error in case a suffix / is present and the deepest node is not a directory
• print an error in case the node does not exist

Add a command ls to list the names of nodes in the root folder:

% acdcli ls
AAA/
Something/
abc/
zzz/
Afile.txt
Bfile.txt
anotherfile.txt
zfile.txt

The output should:

• append / to folder names
• list folders first
• be sorted by name (uppercase before lowercase)
• list all nodes, not just the first paginated output