Does not work as expected using CLI-tools (wget/curl/npm)

[Former user]

When downloading the file via a browser, everything works as expected, but on the commandline, the archive unpacks all files into current working directory, not a new directory. This trips up npm install with a ENOTDIR-error. The md5 sum is equal for CLI and browser-downloads, wich makes me suspect that the unarchiver on my Mac makes a folder automatically, but the tar-command does not behave this way.

[Bryan Turner]

Your suspicion is correct, anonymous. What you're seeing is the behavior of downloading archive and opening it from your browser. It unpacks the files into a directory named after the archive they were extracted from. That directory doesn't actually exist in the archive, however, so commands like unzip and tar xzf won't do it.

When 2.1 is released on Marketplace, for commands like wget or curl, you can add ?prefix=whatever to your URL and all the archive entries will be put in a whatever subdirectory. When opening from your browser, you'll end up with a directory named after the archive containing a whatever directory that then contains the contents, but from unzip and tar xzf you'll end up with a whatever containing the contents.