On 11 November 2010 05:33, Phil Steitz <phil.ste...@gmail.com> wrote: > > > > > On Nov 10, 2010, at 2:07 PM, Gilles Sadowski <gil...@harfang.homelinux.org> > wrote: > >> Hi Luc. >> >>>> Do we need a "task" issue in order to delete all code deprecated in 2.2 and >>>> before? >>> >>> I think we need several different issues, depending on broad topics >>> (exceptions, optimization, ODE, linear algebra ...), and most >>> importantly several commits. I guess this will not be as smooth as we >>> expect, so handling this in a few chunks would be better for us and for >>> the few users who use directly the repository. >>> >>> This would also avoid only one developers has to support all the burden >>> by himself, we can share the tasks. >> >> Of course, I didn't mean that I would remove all deprecated at one fell >> swoop. Several commits are in order, but we could create a single issue >> (where anyone dealing with some part of the work would write in which >> revision the deletion has occurred). My question was rather because an >> alternative option is to not create a JIRA report at all, as it is pretty >> obvious what the issue is with deprecated code :-). >> >> One issue for everything, one per package, none? > > I would say divide it up logically somehow so that it all ends up documented > in manageable chunks in JIRA and the changelog. Some depredations are small > and isolated - no problem just committing with a note in the changelog. The > more complex changes involving removal of entire classes should be documented > in individual issues.
Depredations? Is that what they mean by software rot? :-) +1 otherwise. If we want to keep the JIRAs together, maybe create a main JIRA and attach child tasks to it for the work to be done. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
