diff options
| author | Boris Kolpackov <boris@codesynthesis.com> | 2024-11-20 09:27:21 +0200 |
|---|---|---|
| committer | Boris Kolpackov <boris@codesynthesis.com> | 2024-11-20 11:18:52 +0200 |
| commit | 096362fa82e7ab390078089e4dbbf04da66c79e0 (patch) | |
| tree | ad0a5c2c43ef276eddafb6a4aae51ab2d7b59977 /libbuild2/file.cxx | |
| parent | aec6628e8c1f0ebdb87767cbf804b01504d82459 (diff) | |
Handle config.import.*.<name>.{liba,libs} in cc::search_library()
In particular, this provides support for using prefixed/suffixed installed
libraries (GH issue #449). Note that the current implementation only supports
immediate import but it feels like this limitation can be removed with some
effort in the future.
Diffstat (limited to 'libbuild2/file.cxx')
| -rw-r--r-- | libbuild2/file.cxx | 11 |
1 files changed, 7 insertions, 4 deletions
diff --git a/libbuild2/file.cxx b/libbuild2/file.cxx index f834d8c..8e3b2a4 100644 --- a/libbuild2/file.cxx +++ b/libbuild2/file.cxx @@ -2268,6 +2268,9 @@ namespace build2 // the config.<proj>.<name> variable). For backwards-compatibility // reasons, it takes precedence over config.import. // + // Note also that phase 2 import may handle these imports in an ad hoc + // manner (see cc::search_library() for an example). + // // Note: see import phase 2 diagnostics if changing anything here. // // @@ How will this work for snake-case targets, say libs{build2-foo}? @@ -2364,10 +2367,10 @@ namespace build2 // // config.import.build2.b=b-boot // - // @@ Maybe we should still complete it if it's not simple? After - // all, this is a path, do we want interpretations other than - // relative to CWD? Maybe we do, who knows. Doesn't seem to - // harm anything at the moment. + // Maybe we should still complete it if it's not simple? After + // all, this is a path, do we want interpretations other than + // relative to CWD? Maybe we do, who knows. Doesn't seem to harm + // anything at the moment. Yes, we do, see cc::search_library(). // // Why not call import phase 2 directly here? Well, one good // reason would be to allow for rule-specific import resolution. |