> I believe so. Something so that it tests to see if the header is bad, and if > it is, it remove it and tries (at least once, maybe more) to get the header again. > ok this is in the newest daily. test it out. > I don't really know about what sort of tests to do. > I do know that checking to see if the header.info file is 0 is not a good > test. We have several 'mini-repositories', and many of them don't have > anything in them, so we have blank header.info files. But if I am thinking > correctly, there shouldn't be a header file that is 0. hmmm. If a repositories has no files is it really a repository? <shrug> Shouldn't be too hard to read one line see if it has meaning and do stuff about it. I'll look at the header.info grab again - see what trick I can come up with to check them. Probably nothing difficult. Though, a zero length header would be a surprise - maybe an errorlog at level 1 then ignore that repository? -sv
