On 6/5/06, Andrew Pollock <apollock@debian.org> wrote:
> The issue is that I uploaded -3; it was not a QA upload. My changes in > -3 seem to have been omitted from the QA upload of -4. See (far) below > for the changelog from my -3 upload. > Sure, I read the changelog before composing my previous email. My point remains: when I do a QA upload, I take what's currently in unstable. The only possibility, (and it's quite possible, given that I appear to have uploaded this package twice) is I downloaded what was in unstable at some point between you making your upload and me making mine. If this is the case, then I am entirely at fault, and need to stop working in a vacuum, and apologise profusely.
Sorry, Andrew. I didn't mean to imply that you had reverted the changes intentionally. I only meant to point out that that was the effect, when Martin was wondering why the bug reappeared. There appears to have been some sort of race condition here. Although, if we had both used -2 as a starting point and both uploaded a -3, I don't see how this would not have been caught by the FTP masters, and how it would have resulted in a -4 version. My best guess is that your -3 upload was rejected (due to -3 already exisiting in the archive) which caused you to upload -4 some time later. You started with -3 (which you didn't realise was your rejected -3 and not my -3) made your -4 changes, and uploaded it. Cheers, Shaun