On Mon, 2009-02-23 at 09:53 +0100, Andreas Tille wrote: > These "new" versions are only release candidates (0.4-rc\d) which should > not be packaged (as I agreed with upstream). So I might live with these > false alarms for the next couple of weeks but I would like to fix the > watch file properly. So I tried: > > > http://www.gnumed.de/downloads/client/([\d\.]+)/GNUmed-client\.([\d\.]+)\.tgz > > but this does not even detects the recent version (0.3.10). My guess is > that uscan is at first seeking for the directory with the highest version > ( http://www.gnumed.de/downloads/client/0.4 ) and then parses for the > tarball which matches GNUmed-client\.([\d\.]+)\.tgz - but there is no such > file (these are only in ( http://www.gnumed.de/downloads/client/0.3 ).
No need to guess. :-) uscan(1) agrees that the above is what happens; admittedly it took me a couple of minutes to double-check. (Patches for making uscan(1) more accessible are always welcome :-s) > As you can read in the thread starting at > http://lists.debian.org/debian-qa/2009/02/msg00105.html > I discussed this problem on debian-qa and Paul Wise actually found a > workyround for this problem but we agreed that it would be a nice feature > of uscan to parse lower verisoned directories of the scan of the latest > might have failed. I believe this boils down to basically the same request as #375138, which has been marked wontfix for a couple of years. I haven't merged them yet as I'm undecided as to whether I agree with the original wontfix - opinions welcome, particularly from other devscripts maintainers. Regards, Adam -- To unsubscribe, send mail to [email protected].
