Projects / Parallel URL fetcher

Parallel URL fetcher

puf is a tool that can be used to download single files or mirror entire servers. It is similar to GNU wget (and has a partly-compatible command line), but has the ability to do many downloads in parallel. This is very useful if you have a high-bandwidth Internet connection.

Tags
Licenses
Operating Systems
Implementation

Recent releases

  •  11 May 2002 04:56

    Release Notes: This release fixed a format string vulnerability in error output, and an overflow resulting in bad bandwidth averages for very big downloads or very fast networks. Most switches are now URL-based, a -xq switch was added, the -p switch has been renamed to -B (wget-like), and the -ri switch was replaced by -p, -pr, -pr+ & -pr++ (somewhat wget-like). It no longer compiles on systems without snprintf.

    •  11 Sep 2001 18:08

      Release Notes: This release adds -xH, -ib, and -iU switches. The licencing has been clarified in some places.

      •  21 Jan 2001 14:27

        Release Notes: Basic proxy support has been added.

        •  01 Dec 2000 15:49

          Release Notes: A switch for source address binding was added, and a spec file for building RPMs was created.

          •  26 Oct 2000 15:26

            Release Notes: A bug fix was made for a buffer overflow which caused a segfault when an HTTP server sent a very long header. The default for -ll was reduced to 10.

            Recent comments

            13 Aug 2002 13:35 oswaldbuddenhagen

            Re: Behavior

            > 403 means "access denied".

            oh, really? *surprise surprise*


            > Access modes don't change automatically or randomly.

            explain this to the servers that have a different opinion.

            13 Aug 2002 13:14 Bisqwit

            Re: Behavior

            > % (does 403 reply really mean "try
            > again until it works"?).
            >
            > yes, 403 obviously sometimes is a
            > transient error - otherwise i hadn't
            > done it that way.

            403 means "access denied".
            Access modes don't change automatically or randomly.

            15 Mar 2002 07:32 oswaldbuddenhagen

            Re: Behavior

            > this looks like an ideal DoS tool.

            there are more effective tools for such purposes ...


            > It seems to keep retrying (at least a couple of times) after it has already been banned for flooding
            > (does 403 reply really mean "try again until it works"?).

            yes, 403 obviously sometimes is a transient error - otherwise i hadn't done it that way.
            the retry logic should be a lot smarter and server-based, not url-based, as it is currently.
            also, it should be possible to specify a per-server connection limit, not only a global one. i have already started working in this direction, but the usual permanent lack of time prevents me from finishing it any time soon.

            14 Mar 2002 04:55 Bisqwit

            Behavior
            From the webserver administrator point of view, this looks like an ideal DoS tool. It seems to keep retrying (at least a couple of times) after it has already been banned for flooding (does 403 reply really mean "try again until it works"?).

            Screenshot

            Project Spotlight

            OpenStack4j

            A Fluent OpenStack client API for Java.

            Screenshot

            Project Spotlight

            TurnKey TWiki Appliance

            A TWiki appliance that is easy to use and lightweight.