On Mon, Aug 29, 2016 at 02:52:45PM +0200, Tomasz Nitecki wrote: > tags 835653 + pending > thanks > > > Hey Nicolas, > > On 28/08/16 17:31, Nicolas Braud-Santoni wrote: > > I wrote up a quick patch that takes into account the package names > > when sorting (or bug id, in the case of RFSes), then added the > > removal time (in the case of autorm) and cleaned up a bit the > > surrounding code by using `sort_by` rather than `sort { |a,b| expr(a) > > <=> expr(b) }`. > > Thanks for the patches! :) > > I've applied both 0001 and 0002, so #835653 should be resolved with the > upcoming upload. > > However, I do agree with Paul about making the sort by removal date > feature (patch 0003) optional.
Note that patch 0003 doesn't actually change anything (the order of the sort fields is wrong). How has this patch been tested? > I'm also wondering if sorting by removal > date as the primary sorting field, wouldn't be more useful in this case > (so that the packages that are going to be removed first are at the > top)? What do you think? I was actually about to suggest something similar, since this part of "how-can-i-help --old --all" makes most sense when it tells you what is going to be removed soon. Paul, why are you diffing the daily output of --old? Diffing the --old output should be roughly equivalent to not using --old. Not sure whether it is related to what Paul is doing, but I would call it a bug that how-can-i-help (without --old) lists a package under "New packages going to be removed from Debian 'testing'" each time a comment to a bug report moved that removal further into the future. As an example, "how-can-i-help --all" currently lists fpc and all it's rdeps every day again as "new" due to the discussion in #826300. Similar, O/RFA -> ITA changes should also not trigger listing as "new". > Regards, > T. Nitecki cu Adrian -- "Is there not promise of rain?" Ling Tan asked suddenly out of the darkness. There had been need of rain for many days. "Only a promise," Lao Er said. Pearl S. Buck - Dragon Seed