Extract system header search paths from GCC or compatible

12 files changed, 213 insertions, 31 deletions

diff --git a/build2/c/init.cxx b/build2/c/init.cxx
index 45097bb..e473e58 100644
--- a/build2/c/init.cxx
+++ b/build2/c/init.cxx
@@ -301,6 +301,9 @@ namespace build2
         cast<dir_paths>         (rs[cm.x_sys_lib_dirs]),
         cast<dir_paths>         (rs[cm.x_sys_inc_dirs]),
 
+        cm.sys_lib_dirs_extra,
+        cm.sys_inc_dirs_extra,
+
         c::static_type,
         nullptr,        // No C modules yet.
         hdr,
diff --git a/build2/cc/common.hxx b/build2/cc/common.hxx
index 3c4b693..d4f8bd2 100644
--- a/build2/cc/common.hxx
+++ b/build2/cc/common.hxx
@@ -42,7 +42,7 @@ namespace build2
 
       const variable& x_path;         // Compiler process path.
       const variable& x_sys_lib_dirs; // System library search directories.
-      const variable& x_sys_inc_dirs; // Extra header search directories.
+      const variable& x_sys_inc_dirs; // System header search directories.
 
       const variable& x_poptions;
       const variable& x_coptions;
@@ -123,6 +123,9 @@ namespace build2
       const dir_paths& sys_lib_dirs; // x.sys_lib_dirs
       const dir_paths& sys_inc_dirs; // x.sys_inc_dirs
 
+      size_t sys_lib_dirs_extra;     // First extra path (size if none).
+      size_t sys_inc_dirs_extra;     // First extra path (size if none).
+
       const target_type& x_src; // Source target type (c{}, cxx{}).
       const target_type* x_mod; // Module target type (mxx{}), if any.
 
@@ -164,6 +167,8 @@ namespace build2
             bool fs,
             const dir_paths& sld,
             const dir_paths& sid,
+            size_t sle,
+            size_t sie,
             const target_type& src,
             const target_type* mod,
             const target_type* const* hdr,
@@ -179,6 +184,7 @@ namespace build2
             modules (fm),
             symexport (fs),
             sys_lib_dirs (sld), sys_inc_dirs (sid),
+            sys_lib_dirs_extra (sle), sys_inc_dirs_extra (sie),
             x_src (src), x_mod (mod), x_hdr (hdr), x_inc (inc) {}
     };
 
diff --git a/build2/cc/compile.cxx b/build2/cc/compile.cxx
index c1f4e08..639d3e2 100644
--- a/build2/cc/compile.cxx
+++ b/build2/cc/compile.cxx
@@ -760,8 +760,10 @@ namespace build2
 
             // Extra system header dirs (last).
             //
-            for (const dir_path& d: sys_inc_dirs)
-              cs.append (d.string ());
+            assert (sys_inc_dirs_extra <= sys_inc_dirs.size ());
+            for (auto i (sys_inc_dirs.begin () + sys_inc_dirs_extra);
+                   i != sys_inc_dirs.end (); ++i)
+              cs.append (i->string ());
           }
 
           hash_options (cs, t, c_coptions);
@@ -1709,10 +1711,12 @@ namespace build2
 
           // Extra system header dirs (last).
           //
-          for (const dir_path& d: sys_inc_dirs)
+          assert (sys_inc_dirs_extra <= sys_inc_dirs.size ());
+          for (auto i (sys_inc_dirs.begin () + sys_inc_dirs_extra);
+               i != sys_inc_dirs.end (); ++i)
           {
             args.push_back ("-I");
-            args.push_back (d.string ().c_str ());
+            args.push_back (i->string ().c_str ());
           }
 
           if (md.symexport)
@@ -2670,10 +2674,12 @@ namespace build2
           append_options (args, t, c_poptions);
           append_options (args, t, x_poptions);
 
-          for (const dir_path& d: sys_inc_dirs)
+          assert (sys_inc_dirs_extra <= sys_inc_dirs.size ());
+          for (auto i (sys_inc_dirs.begin () + sys_inc_dirs_extra);
+               i != sys_inc_dirs.end (); ++i)
           {
             args.push_back ("-I");
-            args.push_back (d.string ().c_str ());
+            args.push_back (i->string ().c_str ());
           }
 
           if (md.symexport)
@@ -3879,10 +3885,12 @@ namespace build2
 
         // Extra system header dirs (last).
         //
-        for (const dir_path& d: sys_inc_dirs)
+        assert (sys_inc_dirs_extra <= sys_inc_dirs.size ());
+        for (auto i (sys_inc_dirs.begin () + sys_inc_dirs_extra);
+             i != sys_inc_dirs.end (); ++i)
         {
           args.push_back ("-I");
-          args.push_back (d.string ().c_str ());
+          args.push_back (i->string ().c_str ());
         }
 
         if (md.symexport)
diff --git a/build2/cc/gcc.cxx b/build2/cc/gcc.cxx
index 7f52a91..64c78ca 100644
--- a/build2/cc/gcc.cxx
+++ b/build2/cc/gcc.cxx
@@ -24,6 +24,133 @@ namespace build2
   {
     using namespace bin;
 
+    // Extract system header search paths from GCC (gcc/g++) or compatible
+    // (Clang, Intel) using the -v -E </dev/null method.
+    //
+    dir_paths config_module::
+    gcc_header_search_paths (process_path& xc, scope& rs) const
+    {
+      dir_paths r;
+
+      cstrings args;
+      string std; // Storage.
+
+      args.push_back (xc.recall_string ());
+      append_options (args, rs, c_coptions);
+      append_options (args, rs, x_coptions);
+      append_options (args, tstd);
+
+      // Compile as.
+      //
+      auto langopt = [this] () -> const char*
+      {
+        switch (x_lang)
+        {
+        case lang::c:   return "c";
+        case lang::cxx: return "c++";
+        }
+
+        assert (false); // Can't get here.
+        return nullptr;
+      };
+
+      args.push_back ("-x");
+      args.push_back (langopt ());
+      args.push_back ("-v");
+      args.push_back ("-E");
+      args.push_back ("-");
+      args.push_back (nullptr);
+
+      if (verb >= 3)
+        print_process (args);
+
+      try
+      {
+        // Open pipe to stderr, redirect stdin and stdout to /dev/null.
+        //
+        process pr (xc, args.data (), -2, -2, -1);
+
+        try
+        {
+          ifdstream is (
+            move (pr.in_efd), fdstream_mode::skip, ifdstream::badbit);
+
+          // Normally the system header paths appear between the following
+          // lines:
+          //
+          // #include <...> search starts here:
+          // End of search list.
+          //
+          // The exact text depends on the current locale. What we can rely on
+          // is the unique "#include <" prefix of the "opening" line and the
+          // fact that the paths are indented with a single space character,
+          // unlike the "closing" line.
+          //
+          // Note that on Mac OS we will also see some framework paths among
+          // system header paths, followed with a comment. For example:
+          //
+          //  /Library/Frameworks (framework directory)
+          //
+          // For now we ignore framework paths and to filter them out we will
+          // only consider valid paths to existing directories, skipping those
+          // which we fail to stat or normalize.
+          //
+          string s;
+          for (bool found (false); getline (is, s); )
+          {
+            if (!found)
+              found = s.compare (0, 10, "#include <") == 0;
+            else
+            {
+              if (s[0] != ' ')
+                break;
+
+              try
+              {
+                dir_path d (s, 1, s.size () - 1);
+
+                // @@ Pass true as the ignore_error argument for exists(), when
+                //    supported.
+                //
+                if (d.absolute () && exists (d) &&
+                    find (r.begin (), r.end (), d.normalize ()) == r.end ())
+                  r.emplace_back (move (d));
+              }
+              catch (const invalid_path&) {}
+            }
+          }
+
+          is.close (); // Don't block.
+
+          if (!pr.wait ())
+            throw failed ();
+        }
+        catch (const io_error&)
+        {
+          pr.wait ();
+          fail << "error reading " << x_lang << " compiler -v -E output";
+        }
+      }
+      catch (const process_error& e)
+      {
+        error << "unable to execute " << args[0] << ": " << e;
+
+        if (e.child)
+          exit (1);
+
+        throw failed ();
+      }
+
+      // It's highly unlikely not to have any system directories. More likely
+      // we misinterpreted the compiler output.
+      //
+      if (r.empty ())
+        fail << "unable to extract " << x_lang << " compiler system header "
+             << "search paths";
+
+      return r;
+    }
+
     // Extract system library search paths from GCC (gcc/g++) or compatible
     // (Clang, Intel) using the -print-search-dirs option.
     //
@@ -125,8 +252,10 @@ namespace build2
       //
       for (string::size_type b (0);; e = l.find (d, (b = e + 1)))
       {
-        r.emplace_back (l, b, (e != string::npos ? e - b : e));
-        r.back ().normalize ();
+        dir_path d (l, b, (e != string::npos ? e - b : e));
+
+        if (find (r.begin (), r.end (), d.normalize ()) == r.end ())
+          r.emplace_back (move (d));
 
         if (e == string::npos)
           break;
diff --git a/build2/cc/module.cxx b/build2/cc/module.cxx
index 0245def..141f3c8 100644
--- a/build2/cc/module.cxx
+++ b/build2/cc/module.cxx
@@ -160,27 +160,39 @@ namespace build2
       dir_paths inc_dirs;
 
       if (ci.id.value () == compiler_id::msvc)
+      {
         lib_dirs = msvc_library_search_paths (ci.path, rs);
+        inc_dirs = msvc_header_search_paths (ci.path, rs);
+      }
       else
       {
         lib_dirs = gcc_library_search_paths (ci.path, rs);
+        inc_dirs = gcc_header_search_paths (ci.path, rs);
+      }
+
+      sys_lib_dirs_extra = lib_dirs.size ();
+      sys_inc_dirs_extra = inc_dirs.size ();
 
 #ifndef _WIN32
-        // Many platforms don't search in /usr/local/lib by default (but do
-        // for headers in /usr/local/include). So add it as the last option.
-        //
-        lib_dirs.push_back (dir_path ("/usr/local/lib"));
+      // Many platforms don't search in /usr/local/lib by default (but do
+      // for headers in /usr/local/include). So add it as the last option.
+      //
+      {
+        dir_path d ("/usr/local/lib");
+        if (find (lib_dirs.begin (), lib_dirs.end (), d) == lib_dirs.end ())
+          lib_dirs.emplace_back (move (d));
+      }
 
-        // FreeBSD is at least consistent: it searches in neither. Quoting its
-        // wiki: "FreeBSD can't even find libraries that it installed." So
-        // let's help it a bit. Note that we don't add it for all the
-        // platforms for good measure because this will actually appear as
-        // usually unnecessary noise on the command line.
-        //
-        if (tt.system == "freebsd")
-          inc_dirs.push_back (dir_path ("/usr/local/include"));
-#endif
+      // FreeBSD is at least consistent: it searches in neither. Quoting its
+      // wiki: "FreeBSD can't even find libraries that it installed." So let's
+      // help it a bit.
+      //
+      {
+        dir_path d ("/usr/local/include");
+        if (find (inc_dirs.begin (), inc_dirs.end (), d) == inc_dirs.end ())
+          inc_dirs.emplace_back (move (d));
       }
+#endif
 
       // If this is a new value (e.g., we are configuring), then print the
       // report at verbosity level 2 and up (-v).
diff --git a/build2/cc/module.hxx b/build2/cc/module.hxx
index 08f573b..928e16c 100644
--- a/build2/cc/module.hxx
+++ b/build2/cc/module.hxx
@@ -49,13 +49,22 @@ namespace build2
       translate_std (const compiler_info&, scope&, const string*) const = 0;
 
       strings tstd;
+      size_t sys_lib_dirs_extra; // First extra path (size if none).
+      size_t sys_inc_dirs_extra; // First extra path (size if none).
+
       compiler_info ci; // Note: some members are moved from.
 
     private:
       dir_paths
+      gcc_header_search_paths (process_path&, scope&) const; // gcc.cxx
+
+      dir_paths
       gcc_library_search_paths (process_path&, scope&) const; // gcc.cxx
 
       dir_paths
+      msvc_header_search_paths (process_path&, scope&) const; // msvc.cxx
+
+      dir_paths
       msvc_library_search_paths (process_path&, scope&) const; // msvc.cxx
 
     private:
diff --git a/build2/cc/msvc.cxx b/build2/cc/msvc.cxx
index b7b5828..aa4095e 100644
--- a/build2/cc/msvc.cxx
+++ b/build2/cc/msvc.cxx
@@ -83,6 +83,25 @@ namespace build2
       }
     }
 
+    // Extract system header search paths from MSVC.
+    //
+    dir_paths config_module::
+    msvc_header_search_paths (process_path&, scope&) const
+    {
+      // The compiler doesn't seem to have any built-in paths and all of them
+      // come from the INCLUDE environment variable.
+
+      // @@ VC: how are we going to do this? E.g., cl-14 does this internally.
+      //    cl.exe /Be prints INCLUDE.
+      //
+      //    Should we actually bother? INCLUDE is normally used for system
+      //    headers and its highly unlikely we will see an imported library
+      //    that lists one of those directories in pkg-config Cflags value.
+      //    Let's wait and see.
+      //
+      return dir_paths ();
+     }
+
     // Extract system library search paths from MSVC.
     //
     dir_paths config_module::
diff --git a/build2/cc/pkgconfig.cxx b/build2/cc/pkgconfig.cxx
index f920dbd..a9e3a34 100644
--- a/build2/cc/pkgconfig.cxx
+++ b/build2/cc/pkgconfig.cxx
@@ -1057,9 +1057,6 @@ namespace build2
       //
       pkgconf& ipc (sp.empty () ? apc : spc); // Interface package info.
 
-      // @@ Currently sys_inc_dirs doesn't contain the compiler-extracted
-      //    paths.
-      //
       bool pa (at != nullptr && !ap.empty ());
       if (pa || sp.empty ())
         apc = pkgconf (ap, pc_dirs, sys_lib_dirs, sys_inc_dirs);
diff --git a/build2/config/module.hxx b/build2/config/module.hxx
index d997802..1f48181 100644
--- a/build2/config/module.hxx
+++ b/build2/config/module.hxx
@@ -6,7 +6,6 @@
 #define BUILD2_CONFIG_MODULE_HXX
 
 #include <map>
-#include <algorithm> // find_if()
 
 #include <libbutl/prefix-map.hxx>
 
diff --git a/build2/cxx/init.cxx b/build2/cxx/init.cxx
index ca75de3..25dd7da 100644
--- a/build2/cxx/init.cxx
+++ b/build2/cxx/init.cxx
@@ -511,6 +511,9 @@ namespace build2
         cast<dir_paths> (rs[cm.x_sys_lib_dirs]),
         cast<dir_paths> (rs[cm.x_sys_inc_dirs]),
 
+        cm.sys_lib_dirs_extra,
+        cm.sys_inc_dirs_extra,
+
         cxx::static_type,
         modules ? &mxx::static_type : nullptr,
         hdr,
diff --git a/build2/test/script/regex.cxx b/build2/test/script/regex.cxx
index 38ac096..fe8bcb0 100644
--- a/build2/test/script/regex.cxx
+++ b/build2/test/script/regex.cxx
@@ -4,8 +4,6 @@
 
 #include <build2/test/script/regex.hxx>
 
-#include <algorithm> // copy(), copy_backward()
-
 using namespace std;
 
 namespace build2
diff --git a/build2/test/script/script.cxx b/build2/test/script/script.cxx
index 7ddd63a..eb366c1 100644
--- a/build2/test/script/script.cxx
+++ b/build2/test/script/script.cxx
@@ -5,7 +5,6 @@
 #include <build2/test/script/script.hxx>
 
 #include <sstream>
-#include <algorithm> // find()
 
 #include <build2/target.hxx>
 #include <build2/algorithm.hxx>