Note: This affects RHEL 3 and RHEL 4 If you pass a timestamp to makestamp.py via --timestamp, that variable is set as a string in the script. This causes the following to fail: f.write("%f\n" % data["timestamp"]) Here's a patch: --- makestamp.py 2005-07-21 14:50:04.480315447 -0400 +++ /usr/lib/anaconda-runtime/makestamp.py 2005-04-27 02:02:55.000000000 -0400 @@ -51,8 +51,6 @@ if data["timestamp"] is None: print >> sys.stderr, "timestamp not specified; using the current time" data["timestamp"] = time.time() -else: - data["timestamp"] = float(data["timestamp"]) if data["releasestr"] is None: print "What should be the release name associated with this disc?"
Isn't this diff backwards? :) Also, how important is this for rhel[34] vs just doing it on head?
I don't think the diff is backwards, but it's also past midnight so I'm probably not thinking all that clearly. ;) This isn't a high priority bug as there are workarounds (namely writing the stamp by hand instead of using makestamp.py).
Applied to HEAD