diff options
| author | Boris Kolpackov <boris@codesynthesis.com> | 2016-10-28 11:20:52 +0200 |
|---|---|---|
| committer | Boris Kolpackov <boris@codesynthesis.com> | 2016-11-04 09:26:36 +0200 |
| commit | 39a0cc21f948b251672e1f22d393c33c57a4b649 (patch) | |
| tree | e0dc31a8e5ebe5e36449d96bae27c51039b31732 /build2/test/script/runner.cxx | |
| parent | cd40097447ff2400cb420ec973c16dadd26e6cda (diff) | |
Minor testscript cleanups
Diffstat (limited to 'build2/test/script/runner.cxx')
| -rw-r--r-- | build2/test/script/runner.cxx | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/build2/test/script/runner.cxx b/build2/test/script/runner.cxx index 622696e..8c52288 100644 --- a/build2/test/script/runner.cxx +++ b/build2/test/script/runner.cxx @@ -202,9 +202,9 @@ namespace build2 // // Note that we operate with normalized paths here. // - // @@ What if to make cleanup insensitive to the registration order, - // and sorting paths properly prior removal? Would it produce any - // testing flaws? + // + // @@ I think we should weed duplicates on registration. And just do + // linear search in vector since we don't expect many cleanups. // set<path> rp; for (auto& p: reverse_iterate (sp.cleanups)) |