achilleasa | any ideas why the multiwatcher (core/multiwatcher/types.go) seems to track both Ports and PortRanges? Is it safe to drop Ports and emulate a PortRange -> []params.Port operation for legacy API calls? | 08:37 |
---|---|---|
stickupkid | achilleasa, what happens if you have holes in the PortRange, is that what Ports can do differently? | 08:48 |
achilleasa | stickupkid: we don't seem to be storing ports ATM and everything ports-related has deprecation warnings (see core/network/ports.go) | 08:50 |
achilleasa | I am considering to introduce extra functionality (in a separate PR though) to do port slicing, e.g. open 100-200/tcp, close 150/tcp; result: 100-149/tcp, 151-200/tcp | 08:51 |
stickupkid | achilleasa, seems logical | 08:51 |
achilleasa | plus all the allwatcher bits completely ignore subnets which is a pain | 08:52 |
stickupkid | manadart_, achilleasa should we be using API results for the CLI, we have done in the past, wondering if we should in the future? | 10:17 |
stickupkid | I'm on about this stuff https://github.com/juju/juju/blob/develop/api/charmhub/data.go#L107-L158 | 10:18 |
stickupkid | the issue I have with it, is that we have to encode yaml tags on to types for view only scenarios | 10:21 |
manadart_ | stickupkid: I guess it's less worse if the lib is ours, but technically we should have display types. | 10:22 |
stickupkid | manadart_, exactly my thoughts, but no where in juju does this | 10:22 |
manadart_ | stickupkid: Isn't there generation/branch and space stuff that has display types? | 10:23 |
stickupkid | let me check | 10:23 |
stickupkid | if so that's precedence enough for me | 10:24 |
stickupkid | then we have vertical boundaries and nice code separation | 10:24 |
stickupkid | manadart_, pointer? | 10:25 |
manadart_ | See core/model/generation.go. Those are actually transformed prematurely and returned by the API client, but they are for display. | 10:28 |
stickupkid | manadart_, quick ho? | 10:29 |
manadart_ | And cmd/juju/space/show.go | 10:29 |
gnuoy | I'm looking at a charm which provides a service to multiple applications in multiple models via cross model relations. Would I be right in saying there is no way for the charm to say which relations come from the same remote model ? | 14:09 |
stickupkid | hml, charmhub url model-config is ready for review https://github.com/juju/juju/pull/11834 | 16:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!