diff options
author | Boris Kolpackov <boris@codesynthesis.com> | 2020-09-09 07:31:47 +0200 |
---|---|---|
committer | Boris Kolpackov <boris@codesynthesis.com> | 2020-09-09 07:31:47 +0200 |
commit | 847793d31b2ee057dd9f069c5d354063396aff3b (patch) | |
tree | 198d39a8c888c8e0863ba276080952149c7e5566 /libbuild2 | |
parent | 12d8a2cfcfa10d6de746a919dc995da30850f4ee (diff) |
Expand notes on issues with build system modules that require bootstrap
Diffstat (limited to 'libbuild2')
-rw-r--r-- | libbuild2/module.cxx | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/libbuild2/module.cxx b/libbuild2/module.cxx index bb471a4..11b32cb 100644 --- a/libbuild2/module.cxx +++ b/libbuild2/module.cxx @@ -283,6 +283,12 @@ namespace build2 // places other than the default options file) we would have to propagate // things explicitly. So for now the thinking is that one shouldn't write // such tests except in controlled cases (e.g., module's *-tests). + // + // And another case is the bdep-sync hook which also doesn't have the + // global overrides propagated to it. + // + // It does feel right to propagate global overrides to all the nested + // build system invocations. Maybe we should set an environment variable? module_load_function* r (nullptr); |