Thanks for your detailed reply. Don't worry. I know it's a long ticketWhat is your feature supposed to do now? The thing that you wanted to do was to be able to use the old behavior of why-run mode in order to be able to figure out if there were any queue'd up converge_by blocks, but before the blocks were executed, in order to get a kind of "would_modify_by_action?" check. We completely broke the ability to do that, and as its noted in the ticket its difficult to even predict in the general case that a resource will be updated before you call the action. So what are the conditions under which you expect these before notifications to fire? And why can't you restructure your run_list so that those resources simply come first and are idempotent? Why can't you restructure your run_list to use an after notification?
and this makes it really difficult to review :-)
I reviewed my patches and, in short, the events of the ticket were:
1. My first PR.
2. You recommended me to use why-run to predict updated_by_last_action.
3. Updated to use why-run.
4. You told me that CHEF-3589 broke my path.
5. Rebased and logic changed to work with CHEF-3589.
6. Won't fix.
Since then, it seems that you have not made big changes to why-run
logic. I mean, this code seems to work in current master. I rebased it
here (fixing some small conflicts):
I tested it with some simple recipes and seems to work.
I know my proposal is far from being perfect. But I find it strange
that it was rejected without giving me a chance to improve it. Maybe I
was misguided, but AFAIK nobody wrote me about it :-S
Anyway, based on this code, do you think it appropriate for me to open
a new ticket and propose as a solution? (I can not open the previous
ticket) Or you just don't like my solution and you are looking for
Archive powered by MHonArc 2.6.16.