Hi, I've worked out what seemed to be a correct URL for for rakarrack spec, now that it is accessed via git within the sf website: %global commit a6208406d94a1da978f435605072ee5caefe1491 Source0: http://sourceforge.net/code-snapshots/git/r/ra/rakarrack/git.git/rakarrack-git-%{commit}.zip Which works OK for some time, but then the next day it fails. It seems that the web site creates the archive dynamically, and you can download it multiple times using the above URL, within say 6-20 hours: https://sourceforge.net/p/rakarrack/git/ci/master/tree/ click Download Snapshot, then cancel the auto download, and copy the direct URL. Then next day, I noticed rpmlint is kicking up an error saying source doesn't exist. Indeed attempting to wget the filename fails. But then if I go to the web page again and click Download Snapshot (which creates the archive), now rpmlint succeeds. I couldn't work out how to fit the sourceforce git URLs from: https://fedoraproject.org/wiki/Packaging:SourceURL#Sourceforge.net or https://fedoraproject.org/wiki/Packaging:SourceURL#Git_Hosting_Services to suit. Any ideas for a standard way to handle sf git snapshots ? Is the provided URL (which goes stale quickly, yet is re-generate-able) OK ? -- packaging mailing list packaging@xxxxxxxxxxxxxxxxxxxxxxx http://lists.fedoraproject.org/admin/lists/packaging@xxxxxxxxxxxxxxxxxxxxxxx