aget - A small and performant HTTP/HTTPS/Gemini download tool

Usage:

aget ?

URL/A,TO,BUFSIZE/K/N,MINIMIZEDELAY/S,TIMEOUTSECONDS/K/N,DUMPHEADERS/K,TOSTDOUT/S,ONLYPROGRESS/S,QUIET/S:

URL: The URL to download. If no protocol part is given in the

            URL, it defaults to http://.

TO: Optional destination path. The default is to download to

            the current directory, using the file name from the URL. If

            a drive or directory is specified, it will download there

            instead. If a path with file name is specified, it will

            download to that.

BUFSIZE: Total buffer size in Bytes. This will be rounded to a size

            which makes sense for writing to a file system. Default is

            128kB.

MINIMIZEDELAY: Do not wait until the buffer is filled up before writing.

            Beneficial for minimizing delay, if you are not downloading

            to a normal filesystem, but instead to something that does

            not prefer evenly sized chunks like PIPE:.

TIMEOUTSECONDS: How many seconds to wait until connection times out. The

            default is whatever the connect timeout in the TCP/IP is.

DUMPHEADERS: Dump the headers from the server response to this file.

            The format can be parsed by dos.library/ReadArgs() and

            friends.

TOSTDOUT: Print the downloaded data to the standard output instead of

            downloading it to disk.

ONLYPROGRESS: Only show the progress bar during download.

QUIET: Print no output, except if an error is encountered.

Example:

1> aget https://uhc.a1k.org/UHCTools.lha

Buffer size: 128k

Connected to uhc.a1k.org:443

TLSv1.3 connection is using TLS_CHACHA20_POLY1305_SHA256

Sent GET request!

Server response: 200 OK

Receiving file: 0% - [oooooooooooooooooooooooooooooooo] - 100% (53.23kB)

53.23kB received in 1.07s - 49.42kB/s

Additional usage details:

• A HTTP proxy can be used by setting the ENV-variable http_proxy to the URL

of the proxy.

• Using aget for HTTPS or Gemini requires AmiSSL 4+.

Relevant implementation details:

It uses double buffering and asynchronous writes - downloads from the

network to one buffer at the same time as it writes the other buffer to disk.

This more or less removes the effect seek times and slow disks has on the

download speed (as long as the disk is at least as fast as the network).

This will be most effective if you have a real DMA disk controller, as then

the disk writes taking place at the same time as the network download will

steal a minimum of CPU time from the TCP/IP stack.