From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.5) Gecko/20011012 Description of problem: up2date allowed me to select xinetd-2.3.3-1.i386.rpm anonftp-4.0-9.i386.rpm to update, by default I have download the associated source package selected. When up2date attempts to download the source RPMs for these packages it error's and cancels the entire update process. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Be able to update to the mentioned RPM packages (as I know the source RPM for them does not exist) 2. Try to update to them. 3. On trying to download the source RPM packes up2date errors and cancels the entire process. Actual Results: up2date errors and cancels the entire process. Expected Results: The erro should be logged or message'd boxed - I don't expect the entire process to cancel as it mines I have to download these packages seperatly. Doing this is not a good user experience :-) Additional info: To work around this in my copy, I changed up2date.py such that... def getPackage(pkg, msgCallback = None, progressCallback = None): ... if cfg.readEntry("retrieveSource"): s = getServer() try: pkgStream = up2date.source_package_by_name(getSystemId(), hdr['SOURCERPM']) .... except xmlrpclib.Fault, f: ** if f.faultCode == -17: ** log.log_me("Failed: %s" % f.faultString) ** return 1 ** else: raise CommunicationError(f.faultString) Added the ** lines so that the cannot download error for a source RPM does not cancel the entire process. Works for me!! :-) I'm sure something better can be done with the error.
Well, if you cant download the src rpms the user asked for, I think throwing up an error message is the correct thing to do for now. Ideally we would throw up an error message and have some way to ignore it if we like, but that will have to wait for the next version. The real question is "why are source rpm downloads failing?"
The problem with ignoring missing source packages is: what status code do you return when up2date exits? It would need to be non-zero to be useful for testing. Admittedly, this is only a problem for testers, and a testers problem shouldn't really be made a users problem.
So is it a bug or ain't it a bug?
not a bug... if the download fails: a). up2date didnt do what the user wanted. Thats a failure and shouldnt be ignore. b) It means something is wrong with the server, and should be fixed.