On 04/02/08 at 23:32 +0100, Thijs Kinkhorst wrote: > On Monday 4 February 2008 23:14, Lucas Nussbaum wrote: > > This is a recommended practice, not a mandatory one. And in practice, > > when processing a lot of NMUs, it is much easier to: > > 1. contact the maintainer through the BTS > > 2. immediately upload to DELAYED/n > > 3. if the maintainer answers, cancel/delay some more the upload > > Than to: > > 1. contact the maintainer through the BTS > > 2. wait an "appropriate" amount of time > > 3. upload without going through DELAYED > > > > The result is exactly the same, and, FWIW, during all the NMUs I sponsored > > to fix dash build failures, you were the only one to complain. > > Although this may be technically very sound and yield the same result, I do > agree that the signal sent by uploading to DELAYED/2 right away is quite a > different one from the second scenario. > > Not to discredit your very useful work, but I hardly see any reason why this > bug needed to be fixed two days later already. Of all the choices of the > delayed queue, 2 instead of 10 suggests some real urgency which I cannot find > in this bug.
2 instead of 10 simply means that there was no point in waiting 8 more days for a trivial fix. Uploads to DELAYED can be delayed some more, and I always make sure that I will be around at the end of the delay, so, if the maintainer needs more time, I can delay the upload some more (which happend in the case of dancer-services). FWIW, to improve the signal sent to the maintainer a bit, I changed the "template" I use when sponsoring NMUs to: | Hi, | | I have sponsored this NMU, and uploaded it to DELAYED/3. Don't hesitate | to tell me if you want me to delay it some more, in case you want to do | the upload yourself. | | Thank you, > Uploading to the DELAYED/10 queue has the exact same effect for you (no need > to come back to it) but allows for much more time, Then why 10, and not 20? I think that it can be safely assumed that (when not in VAC) 99% of active DDs get a chance to read their mail at least once every 48 hours, and send a simple message like "I'd like to do the upload myself, but don't have time right now, please delay it for a few more days". And if they don't, the consequences aren't terrible: a fix will be uploaded to unstable, and they can always make an upload later. That only causes a bit more stress on our infrastructure. Maybe we should write a DEP about "Recommended practices for NMU" to clarify all this. Is someone interested in driving this with me? -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
