-
Notifications
You must be signed in to change notification settings - Fork 325
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Provide a way to change the behavior when picking up default TestAdapters #2580
Comments
It might be worth thinking about multiple scenarios. It would be nice if this fits well with enabling just adapters that are next to the source, and possibly allowing additional adapters to take path to the dll not just to directory. This would hopefully improve the perf as well if there are any data collected on that already @Sanan07 ? I think flags choice for how adapters are picked up would be nice (with some nicer naming), where Default is the default:
By also adding the ability to provide files directly to the additional test adapters, we should be able to narrow down the discovery much more than we are now. Then we need to look at how easy it is to enable this though, and how much time it can save. |
/cc: @jakubch1 |
Consider having separated configs for test adapters and data collectors |
It would be good to be able to specify order. For example
|
If the |
@anirudhsanthiar this behavior we'll be inside in our next preview release this Monday.
Unfortunately this affects Test Platform very fundamentally we'll aim to provide this support in our next major release. |
lib.dll -> MD5
|
@MarcoRossignoli if that is the only way to specify the version then using md5 seems to be terrible for maintenance and upgrades. If that is only a mode for choosing a very specific dll and we have another mode in place that checks versions and chooses newer then I agree. But it also sound a bit unnecessary, and as an additional thing we would need to maintain together with the other mechanism that leverages versions. |
it's to fix "the unexpected scenario" where the order is not enough. |
Ah so that would be exclude the dll with this MD5, rather than use that dll with MD5? That was not very clear from the PR title. |
I'd say both side the chance to specify include/exclude. |
Also discussing with @jakubch1, the idea to have another "folder" like "FirstChance", where the user put extensions and it will be check for first during the normal loading order. Always for "the unexpected scenario", to avoid the too strict MD5 strategy. |
I fear that having another special place where to put stuff will become obsolete the same way the current folder became obsolete. But I am looking forward to reviewing your plan. :) |
I like the plan of folder ordering, but I think that we need to have a workaround always fixes possible uncovered scenario. |
By "FirstChance" we understand a way that you add specific folder into TestAdapterPath For example in runsettings it could be like this:
The issue is that azure pipelines are adding something to beginning of TestAdapterPath so it would be good to be able to specify order of usage of TestAdapterPath itself. |
cc: @Evangelink |
cc: @AbhitejJohn |
There is an implementation for this in #3380. It doesn't currently support ordering - because order of some adapters are hard set. I allow them to be excluded, or included. A basic implementation is out with 17.1 NuGet package. |
Issue
VSTest automatically picks up all the
*TestAdapter.dll
assemblies present with the assembly tested. We don't offer a way to change this yet. A customer reported on a feedback ticket that this delays their tests for a long time.Proposal
Since TestAdapters and DataCollectors can be in the same assembly, specifying a different load order is not feasible. We'll use the same loading order for the assemblies, and provide a way to exclude them.
AB#1370910
The text was updated successfully, but these errors were encountered: