Welcome to CRUX bug tracking.
FS#684 - Add User-Agent to ck4up HTTP Get
Attached to Project:
CRUX
Opened by Daryl Fonseca-Holt (wyatt) - Sunday, 12 December 2010, 21:07 GMT+2
Last edited by Juergen Daubert (jue) - Tuesday, 14 December 2010, 17:49 GMT+2
Opened by Daryl Fonseca-Holt (wyatt) - Sunday, 12 December 2010, 21:07 GMT+2
Last edited by Juergen Daubert (jue) - Tuesday, 14 December 2010, 17:49 GMT+2
|
Detailssourceforge is now tailoring http based on the browser User-Agent. ck4up isn't getting Linux related information, it's getting Windows (argh, the default). In ck4up.conf I have an item that worked until recently:
@SFfiles@ http://sourceforge.net/projects/@NAME@/files/ @TAR@ @NAME@-.*?\.tar\.[bg]z2? xchm md5 @SFfiles@ @TAR@ I worked up a Ruby noob's patch that gets it working again by supplying the User-Agent from Linux Midori but it has the side-effect of requiring an ending slash (/) on each URL or the HTTP Get fails. The attached patch is for ck4up ver 1.2 rel 1. |
This task depends upon
ck4up.diff
Short of adding some kind of configuration setting of OS for ck4up I don't know of any way to address this in ck4up. I can easily make my own port for ck4up and apply a patch to it.
I used Midori for two reasons: it was one of the shorter User-Agent strings at http://zytrax.com/tech/web/browser_ids.htm and in-case they were counting browser types I wanted Midori to get some credit as Mozilla gets plenty. They may only be check the platform portion of the string anyway so it might be possible to plug ck4up as well. I haven't experimented with that yet.
If we can add something like "User_agent" that way and, that's most important for me, get it working without any other required changes like adding "/" to all URL that are directories, I'm not opposed to accept your patch for a future release.