-
-
Notifications
You must be signed in to change notification settings - Fork 990
Fix MsBuildArgument
s
#2393
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
Fix MsBuildArgument
s
#2393
Conversation
898d67e
to
b766377
Compare
It was added as a fallback, for scenarios where |
Okay, that's what I thought. If that's the only case for it, then we don't need it anymore, as specifying a new output path avoids the issue. |
I've tried that in the past (years ago, it could be outdated), one of the issues I've hit was the path being overwritten by some https://github.com/dotnet/performance is a great example of repo that run into the issue. You can use this project file to tell it to use your local copy of BDN rather than a nuget package for testing and then try run the benchmarks from here |
Cool. I tried that and it worked no problem.
I don't know what you tried, but I tried placing it in the csproj and it didn't work, it only works if I pass it to the cli. And if I only use |
7fa2824
to
8f35e9a
Compare
2a3c9a4
to
b451881
Compare
This PR fixes several issues. Could you PTAL @adamsitnik? |
It seems like @adamsitnik is unavailable at the moment, @timcassell would there be anyone else that can review this? |
@timcassell no concerns from my side, but for this issue, we really need a review from @adamsitnik who is the author of our disassembler. I don't know much about this subsystem, and I could miss some potential problems. If you feel like this change is safe and you properly tested it so there is no actual need for a thorough review, let me know, and I will merge it as is. |
@AndreyAkinshin Thanks. The changes I made to the disassembler were simply to release the pdb file when it's no longer needed. Without doing that, tests were failing when it was trying to overwrite the file. There was a todo there to clean up that caching anyway. I think it's safe. I did put this on the v0.14.0 milestone because it is changing how builds are done and breaking some public APIs, but if you want to pull it into v0.13.13, I'm okay with that. |
I don't mind releasing v0.14.0 as the next version. Could you please add some release notes to this PR then? Just create a file If you have any other breaking changes you would like to include in v0.14.0, don't hesitate to ping me in the corresponding PRs. |
@timcassell could you please provide some additional context in the changelog? There is no need for an extensive description, but it would be awesome to have several sentences describing the actual change so that it's easily understandable for our users who don't have the internal development context. The current description feels too brief and doesn't provide a lot of useful information for the users. |
@timcassell now the release notes look great, thanks! |
@AndreyAkinshin I fixed it up. For future reference, "squash and merge" should've worked, rather than "rebase and merge". |
…dotnet` command. Removed `--no-dependencies` build fallback. Release pdb files when they are no longer needed for disassembly. Added v0.14.0 changelog.
Fixes #1377
I removed build with
--no-dependencies
because it apparently makes the resulting build not work properly when paired with--output
. Looking at the blame history, I couldn't find a PR or issue for why it was added in the first place, only a commit (6ee21bb). Is--no-dependencies
needed for a particular reason @adamsitnik?