NUnit3Templates update - #5634
Conversation
….NET.Test.Sdk on netcorepp2+
21ef170 to
fcd7640
Compare
|
Sorry, this was missed, and I think it is too late for 3.1.100. It will have to go in 3.1.101 if we can get approval. @halex2005 I suggest tagging @dotnet/dotnet-cli in the future. |
|
Why it is late? In my opinion, this PR do not affect on SDK, but increase its quality. It just contains simple string with new version of other package, and all checks have passed btw! |
We already have the presumed final build in hand. The bar is extremely high to reset that. There are factors WAY outside my control on when things can be approved. There's a work back from other dates like when the final VS 16.4 build will be and how long to insert into that, etc. It would still have been tight, but if we had jumped on the PR when it came in, we would have had a good chance to get it in. For that not happening, I've apologized above and I'm apologizing here again. The only excuse I have is that we have a lot of repos/branches/releases going on at the same time. Feel free to ping me directly on future PRs to ensure that they are looked at sooner.
I do not unless there is no choice. There are not enough details to fully understand the bug yet. Can you describe:
I will submit these answers for approval, but it is a long shot. I do not make the decisions here.
We all care deeply about this. Holding a high bar at the end of a release helps that more than it hurts, though individual decisions might sting. If our only bar was that PRs were green and believed to increase quality, we would definitely introduce more last minute regressions and actually lower quality of final release. Any given change can be low risk, but the cumulative risk can still be very real, given enough changes. Out of fairness, we have to apply the same bar to all the changes. |
|
@nguerrera, I understand that you have some timetable and release cycle. I heard that 3.1.100 will be released in the beginning of december (or hopefully in the end of november), it is 10 days to release btw. This definitely will not break anything if you merge it, and you definitely should not retest all the stuff after merge. It just will contains up-to-date package versions. For me, that is just the sanity. Btw, it is ok for me if this PR will have to go in 3.1.101. I just would like to have this PR merged. PS. You said that you have |
Thanks for understanding. :)
As I said, there's a work back from that. 10 days to release is not 10 days to checkin.
Even if that happens, I cannot merge this without approval from a group of people that review late changes and servicing changes. To get that approval, I would need to supply the answers to my questions above, so it would be helpful if you could provide the answers. Also, the same questions would apply for 3.0101 so the answers are needed either way. Once we enter servicing, all changes are subject to this approval. |
No, this is the correct branch for 3.1.101, once 3.1.100 is officially final, then changes for 3.1.101 can go into the branch. |
First of all, nunit dotnet new templates are works correctly. But there is subtle bug when targeting .netcoreapp >= v2 in templates - wrong condition will emit
Customers will update
Yes, there is workaround: customers could install lateset templates with
There is no risks. All will works, but |
|
Thank you. |
This PR updates NUnit3.DotNetNew.Template versions with following fixes (see nunit/dotnet-new-nunit#33):
Microsoft.NET.Test.Sdkdependency to v16.4.0Microsoft.NET.Test.Sdkfor netcoreapp2 and higher