Created attachment 449360 [details] wget log wget --mirror devmanual.gentoo.org --rate-limit=100k Results in an infinite dig. Shouldn't wget notice that the URLs are going to the same page and not attempt to recurse again? Could also be a badly slashed link on the devmanual...
Created attachment 449362 [details] emerge --info
I fixed a double-slash in the devmanual. https://gitweb.gentoo.org/proj/devmanual.git/commit/?id=627ca55670862b7bcab101b9f2d30cd6f467e081 Giving this to base-system to address the wget issue.
double slashes aside, I filed this bug not just to get the devmanual fixed, but also because if wget is saving the file in the same spot either way it should also have caught that it was a duplicate. Anything after the domain name (including port) is a file path and should be treated accordingly.
(In reply to Raymond Jennings from comment #3) that is simply not true. the remote side is free to interpret things however it likes, including treating of double slashes differently. here's a live example. these two URLs do not produce the same page: https://www.gnu.org/software/autoconf/manual/ https://www.gnu.org/software/autoconf//manual/ i don't think there's a bug here. your output didn't show an infinite loop, it showed fetching of the same set of resources. of course, when wget operates in mirror mode, it makes assumptions about the behavior of slashes and files on disk. there isn't a good answer here. either way, this should go upstream if you want to pursue it.
