cylc rose

9 Members
1 Servers

Load older messages

Timestamp Message
23 Jan 2019
09:16:44@matthewrmshin:matrix.orgMatt ShinRedacted or Malformed Event
12:34:06@wxtim:matrix.orgwxtimNasty little wee side issue in rose-macro (rose/macro.py) at line 1637 makes the task fail without anything output to stderr or stdout. Have put a small proposed change in.
29 Jan 2019
22:32:52@matthewrmshin:matrix.orgMatt ShinMy latest attempt in using a Github project to document our 2to3 activities. https://github.com/metomi/rose/projects/6
22:41:48@revilo666:matrix.orgHilary OliverLooks good!
22:49:36@matthewrmshin:matrix.orgMatt Shin The interface seems to have improved slightly compared to when I first tried it, but it can also be because I'm looking at things with a different perspective this time.
31 Jan 2019
21:42:43@dwsutherland:matrix.orgDavid Sutherland changed their profile picture.
21:43:10@dwsutherland:matrix.orgDavid Sutherland changed their profile picture.
22:01:13@dwsutherland:matrix.orgDavid Sutherland changed their profile picture.
22:01:58@dwsutherland:matrix.orgDavid Sutherland removed their profile picture.
22:03:03@dwsutherland:matrix.orgDavid Sutherland set a profile picture.
1 Feb 2019
10:38:00@matthewrmshin:matrix.orgMatt Shin In case nobody noticed, I have started a Github project on cylc/cylc to document the rose suite-run migration to Cylc work.
19 Feb 2019
23:48:19@revilo666:matrix.orgHilary OliverDistracted from cylc-8 this morning by some minor Rose development (we need suite log symlinking at NIWA) ... just submitted a PR: https://github.com/metomi/rose/pull/2297
23:51:47@revilo666:matrix.orgHilary Oliver invited @kinow:matrix.orgBruno P. Kinoshita.
23:52:45@kinow:matrix.orgBruno P. Kinoshita joined the room.
23:53:07@kinow:matrix.orgBruno P. Kinoshita👋 Hello
23:57:19@revilo666:matrix.orgHilary Oliver(Just noticed you hadn't joined this room; I should update you on the aforementioned NIWA HPC cylc problem later...)
20 Feb 2019
00:01:55@kinow:matrix.orgBruno P. KinoshitaI probably missed this one. Will read some of the previous messages (I was reviewing a rose issue by coincidence, about py3)
14:06:01@sadielbartholomew:matrix.orgSadie Bartholomew Hi Hilary and Bruno! You arrived in good time for a question that has just occurred to me. I was reading through an issue just opened in Cylc earlier by a user (#2960), in which it occurred to me that there are certain issues, such as that one, that could be argued to be the domain of either Cylc or Rose, but if course issues are only opened against one codebase. I am not aware of any process that we have to account for this, e.g. to acknowledge that an issue in Cylc concerns only Cylc when used with Rose rather than standalone, or that an issue has been opened on the wrong codebase. In the former case, would a 'Cylc with Rose' or similar issue label be a good idea? And for the latter, there is in the Beta version of GitHub the possibility to transfer an issue to another codebase, with the author's permission - should we make use of this in agreed cases (it has to be enabled though I believe, as the option 'transfer issue' and 'pin issue' appear under 'lock conversation' only for Rose issues and not Cylc ones at the moment, at least when viewing with my own user account. Anyone have thoughts on this (including if perhaps they don't think it important to discuss, in which case fair enough)?
14:57:59@matthewrmshin:matrix.orgMatt Shin

It remains our strategy to separate Cylc and Rose as much as possible so we don't have this problem in the first place. The rose suite-run migration work should solve the biggest part of this. rose task-run and built in applications such as rose_prune should then be the remaining utilities with inter project dependency. In addition, any new web UI based on Cylc 8 architecture will introduce some new inter project dependency.

The current way is to have an issue opened in each project with hyperlinks between each other. We can continue to do that should inter project issues occur in the future.

18:57:58@sadielbartholomew:matrix.orgSadie Bartholomew Sure, thanks for the explanation Matt Shin . That seems sensible.
18 Mar 2019
10:49:34@kinow:matrix.orgBruno P. Kinoshita invited @wxtim2:matrix.orgwxtim2.
10:49:35@wxtim2:matrix.orgwxtim2 joined the room.
16:39:07@wxtim2:matrix.orgwxtim2 I'm getting underway on RSR. 8 tests failed out of the box. Now down to six.
17:33:30@matthewrmshin:matrix.orgMatt Shin
In reply to @wxtim2:matrix.org
I'm getting underway on RSR. 8 tests failed out of the box. Now down to six.
Not bad
17:34:27@matthewrmshin:matrix.orgMatt ShinHowever many tests are skipped unless you are running with remote hosts.
19:27:56@wxtim2:matrix.orgwxtim2 True
22:19:13@revilo666:matrix.orgHilary Oliver Which reminds me, we should revisit the idea of getting Bruno P. Kinoshita to demo and explain use of Docker with Cylc (he was experimenting a while back, and I tested his recipe... which worked great) ... so that we can easily test remote hosts (and potentially even with PBS) in a local development environment.
22:21:08@kinow:matrix.orgBruno P. KinoshitaTravis-CI supports Docker too, though I never used that. If we use a simple alpine linux container, it's likely that it could work in Travis-CI too. Our test jobs would take a bit longer, but we can work around that by runnign only once a day, or only for one branch, etc
22:24:20@matthewrmshin:matrix.orgMatt ShinDefinitely. I really want to set up a Docker swarm to simulate a cluster of suite hosts and job hosts, etc.

There are no newer messages yet.

Back to Room List