This was prompted by #1060 and the bugfix for it - there are a number of GCs involved, one for "options" one for "dns options", and whatnot...
13:07 <@plaisthos> we seem to have that sepearete gc for pre push save and then
restore
13:07 <@plaisthos> would probably been better to have that a generic pre push
gc instead of just for dns but that is a different topic
13:10 <@plaisthos> so yeah the patch makes sense but I wish the role of the gc
would be better explained in trhe source code
13:40 <@ordex> plaisthos: that can indeed be refactored. so that GCs are not
per-coponent, but rather "per-lifecycle". but it may entail a
quite large refactoring..
... recording that discussion so we don't forget... definitely not 2.7 material.
This was prompted by #1060 and the bugfix for it - there are a number of GCs involved, one for "options" one for "dns options", and whatnot...
13:07 <@plaisthos> we seem to have that sepearete gc for pre push save and then
restore
13:07 <@plaisthos> would probably been better to have that a generic pre push
gc instead of just for dns but that is a different topic
13:10 <@plaisthos> so yeah the patch makes sense but I wish the role of the gc
would be better explained in trhe source code
13:40 <@ordex> plaisthos: that can indeed be refactored. so that GCs are not
per-coponent, but rather "per-lifecycle". but it may entail a
quite large refactoring..
... recording that discussion so we don't forget... definitely not 2.7 material.