Attention is currently required from: fixeria, osmith, pespin.
1 comment:
Patchset:
IMHO you are adding lots of complexity to regular APIs for a niche use case which should probably be […]
the reason why I like the implicit -1 is exactly because it is so simple and elegantly "docking" to current usage without interfering:
A caller can test for a mix of absent and present counters, setting the expected-absent ones to -1, the expected-present ones to a value >= 0.
The main visible feature of hnb_persistent is the counter, so why not use counter errors to indicate an absent object? Same for all other dynamic ctrgs in all our software.
AFAICT this is the minimal possible patch to achieve this, seems to me your suggestion is the more complex one?
We need others please to give opinions because it seems Pau and I will always disagree on this. Thanks!
To view, visit change 37189. To unsubscribe, or for help writing mail filters, visit settings.