diff options
| author | Dylan DPC <dylan.dpc@gmail.com> | 2020-12-18 00:30:21 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2020-12-18 00:30:21 +0100 |
| commit | f44c2271d073de42e10e46ab76f69f5287b47236 (patch) | |
| tree | 04105ecced9a9f945a4c5872af75ced2b36e35bc /compiler/rustc_codegen_llvm/src | |
| parent | da89dbb41af5fd9f417b30e09451640d68fb50af (diff) | |
| parent | c9fab5008479012ed944a8e542ed2a309ec2a385 (diff) | |
| download | rust-f44c2271d073de42e10e46ab76f69f5287b47236.tar.gz rust-f44c2271d073de42e10e46ab76f69f5287b47236.zip | |
Rollup merge of #80109 - richkadel:llvm-coverage-counters-2.3.0, r=tmandry
Remove redundant and unreliable coverage test results The `coverage-reports` tests still generate counters and JSON reports for inspection, but these files are no longer used in Makefile diffs, to reduce complexity and confusion from unreliable or unexpected test results, especially when maintaining them (i.e., generating `--bless`ed results). The associated `expected_` files for counters and JSON reports have been removed, leaving only the files actually used for testing: the `llvm-cov show` reports. r? `@tmandry` Tyler - as we discussed offline... FYI: `@wesleywiser` `@Swatinem` Arpad, depending on the timing of this PR, it may not affect you, but I'm removing some of the files that produce slightly different results on Windows as they really aren't necessary to validate coverage results.
Diffstat (limited to 'compiler/rustc_codegen_llvm/src')
0 files changed, 0 insertions, 0 deletions