Changes between and of Initial VersionVersion 1Ticket #2424, comment 38


Ignore:
Timestamp:
01/12/21 22:08:19 (9 months ago)
Author:
Glyph
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #2424, comment 38

    initial v1  
    11To summarize my previous comment:
    22
    3  1. This ticket's description, as written, cannot be satisfied. There's no way to run the reactor in a different mode, because a different mode would fundamentally change the meaning of the values returned and accepted by existing interfaces.  I think this is why we've stalled for so long about actually ''doing'' anything about this.
     3 1. This ticket's description, as (previously) written, cannot be satisfied. There's no way to run the reactor in a different mode, because a different mode would fundamentally change the meaning of the values returned and accepted by existing interfaces.  I think this is why we've stalled for so long about actually ''doing'' anything about this.
    44 2. `callLater`'s semantics are tied to wall-clock time and so any compatible replacement (unfortunately) has to preserve that.  Therefore, long term, we might be looking at deprecating `callLater`.
    55 3. Since different applications have different relationships to time, it is not as simple as replacing `callLater` with `callAtTime`; we need ''both'' `callAt` and `callAfter`.