New ebuild for trickle. trickle is a portable lightweight userspace bandwidth shaper. It can run in collaborative mode (together with trickled) or in stand alone mode. trickle works by taking advantage of the unix loader preloading. Essentially it provides, to the application, a new version of the functionality that is required to send and receive data through sockets. It then limits traffic based on delaying the sending and receiving of data over a socket. trickle runs entirely in userspace and does not require root privileges. Reproducible: Always
Created attachment 164149 [details] [Ebuild] trickle-1.06
- The ebuild header is invalid¹ - arch -> ~arch, did you test on all these architectures? - remove src_unpack(), Portage does such trivial cases transparently. - einstall is an ugly hack and only to be used, when emake DESTIDR=${D} install or similar doesn't work. [1] http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=3
Created attachment 164188 [details] Cleaned up ebuild Been a few years since I've played with Gentoo. This ebuild should be better.
trickle 1.07 exists, but it's not announced in the webpage: http://monkey.org/~marius/trickle/trickle-1.07.tar.gz
Before trickle gets into portage, we sure need to fix it a little, 'cause autotools part looks like a total fail — both in 1.06 and 1.07. I've created a few patches to allow it to be compiled on current Gentoo amd64 workstation, but I suppose they can break it somewhere else.
Created attachment 171543 [details, diff] correct in_addr_t search includes trickle's configure looks for in_addr_t in sys/types.h, and on Linux it's in netinet/in.h instead. This causes search to fail and in_addr_t to be redeclared, which in turn causes compilation to fail. This patch adds additional include to configure check.
Created attachment 171546 [details, diff] AM_CFLAGS fix Replaces += with = in AM_CFLAGS declaration. Not sure, whether it's good solution, but it works.
Created attachment 171547 [details, diff] Removal of some ancient hack Removes some ancient hack causing automake to fail. In 1.07 there's note saying that it's probably unnecessary now.
Created attachment 171548 [details, diff] Remove overload_DATA Looks like some predecessor or clone of trickleoverload_DATA, causes 'make install' to fail.
Created attachment 171549 [details] Ebuild with patches
Just tested, and this still works 2 years later, anything in particular keeping it from being merged (aside from upstream having semingly vanished)?
*** Bug 358711 has been marked as a duplicate of this bug. ***
This works, why isnt it in portage?
Created attachment 312223 [details] trickle-1.07 overlay files I just played a little bit with trickle. For anyone who who cares i attached my files to build a working 1.07-overlay-version. Too bad trickle cant limit the upload bandwith from sshd... Cheers! :)
(In reply to Gerrit Helm from comment #14) > Created attachment 312223 [details] > trickle-1.07 overlay files I just tried this out and automake pukes on it now. :-(
https://github.com/mariusaeriksen/trickle may have newer code/Autotools scripts.