Seemed indeed to be a version problem. I downloaded the 1.8.1 sources and was actually allowed to build the thing on my ISP machine! Never thought that would work.
It looks like wget still un-encodes '$' and '*', but not ';'. And perhaps more important it does not encode ':'. Works fine now. Joop > I am trying to download some images from mapquest. These have URL's > with encoded characters. The URL I try to download with wget is > extracted from a html page served by mapquest. > > With the windows WGET (1.8.1b) the url text is passed unaltered to the > web server. This works fine. > > With unix (tried with 1.7 on Linux and BSD) the already encoded > characters like %24, %2A and %3B get turned in '$', '*' and ';'. > Oppositely the ':' is now encoded on unix into %3A... (I do have the > url in "quotes".) As a result the server gives an error. > To be absolutely sure about this re-encoding I verified this by > 'wgetting' from my apache server an looking in the logfile. > > Is there a way to disable this tampering with the URL text by wget 1.7? > > > If it is related to 1.7 - 1.8 version difference, then I am in trouble. > I want to run this on my ISP server over whose version I have no > control. Unless I can find a 1.8 binary for "BSD/OS 4.0 Kernel #7: > i386"... > > Joop
