refactor file filtering

Author: vinniefalco

include/mrdox/Config.hpp

@@ -48,7 +48,7 @@ struct Config
     // Directory where processed files are stored. Links
     // to definition locations will only be generated if
     // the file is in this dir.
-    std::string SourceRoot;
+    std::vector<llvm::SmallString<0>> includePaths;
 
     // URL of repository that hosts code used
     // for links to definition locations.
@@ -61,7 +61,16 @@ struct Config
     Config(Config&&) = delete;
     Config& operator=(Config&&) = delete;
 
-    bool shouldSkipFile(llvm::StringRef filePath) const noexcept;
+    /** Returns true if the file should be skipped.
+
+        If the file is not skipped, then prefixPath
+        is set to the portion of the file path which
+        should be be removed for matching files.
+    */
+    bool
+    filterFile(
+        llvm::StringRef filePath,
+        llvm::SmallVectorImpl<char>& prefixPath) const noexcept;
 
 public:
     struct filter { std::vector<std::string> include, exclude; };

include/mrdox/Visitor.hpp

@@ -26,6 +26,7 @@
 #include <clang/AST/ASTConsumer.h>
 #include <clang/AST/RecursiveASTVisitor.h>
 #include <utility>
+#include <unordered_map>
 
 namespace clang {
 namespace mrdox {
@@ -38,9 +39,18 @@ class Visitor
     : public RecursiveASTVisitor<Visitor>
     , public ASTConsumer
 {
+    struct FileFilter
+    {
+        llvm::SmallString<0> prefix;
+        bool exclude = true;
+    };
+
     bool handleTranslationUnit_ = false;
     tooling::ExecutionContext& exc_;
     Config const& config_;
+    std::unordered_map<
+        clang::SourceLocation::UIntTy,
+        FileFilter> fileFilter_;
 
 public:
     Visitor(

source/lib/Config.cpp

@@ -49,15 +49,6 @@ namespace mrdox {
 Config::
 Config()
 {
-    if (auto p = ::getenv("MRDOX_SOURCE_ROOT"))
-        SourceRoot = p;
-    else
-    {
-        llvm::SmallString<128> SourceRootDir;
-        llvm::sys::fs::current_path(SourceRootDir);
-        SourceRoot = SourceRootDir.c_str();
-    }
-
     if (auto p = ::getenv("MRDOX_REPOSITORY_URL"))
         RepositoryUrl.emplace(p);
     else if (auto q = ::getenv("DRONE_REMOTE_URL"))
@@ -66,10 +57,27 @@ Config()
 
 bool
 Config::
-shouldSkipFile(
-    llvm::StringRef filePath) const noexcept
+filterFile(
+    llvm::StringRef filePath,
+    llvm::SmallVectorImpl<char>& prefixPath) const noexcept
 {
-    return false;
+    namespace path = llvm::sys::path;
+
+    for(auto const s : includePaths)
+    {
+        if(filePath.starts_with(s))
+        {
+            prefixPath = s;
+            if(! path::is_separator(prefixPath.back()))
+            {
+                auto const sep = path::get_separator();
+                prefixPath.insert(prefixPath.end(),
+                    sep.begin(), sep.end());
+            }
+            return false;
+        }
+    }
+    return true;
 }
 
 std::error_code
@@ -78,6 +86,7 @@ load(
     const std::string & name,
     const std::source_location & loc)
 {
+#if 0
     auto ct = llvm::MemoryBuffer::getFile(SourceRoot + "/" + name);
     if (!ct)
         return ct.getError();
@@ -87,7 +96,7 @@ load(
     yin >> *this;
     if ( yin.error() )
         return yin.error();
-
+#endif
     return {};
 }
 

source/lib/Visitor.cpp

@@ -22,6 +22,7 @@
 #include <clang/Index/USRGeneration.h>
 #include <llvm/ADT/StringExtras.h>
 #include <llvm/Support/Error.h>
+#include <llvm/Support/Path.h>
 
 namespace clang {
 namespace mrdox {
@@ -41,26 +42,48 @@ bool
 Visitor::
 mapDecl(T const* D)
 {
+    namespace path = llvm::sys::path;
+
     clang::SourceManager const& sm =
         D->getASTContext().getSourceManager();
-    clang::SourceLocation const loc = D->getLocation();
 
-    if(sm.isInSystemHeader(loc))
+    if(sm.isInSystemHeader(D->getLocation()))
     {
         // skip system header
         return true;
     }
 
-    if (D->getParentFunctionOrMethod())
+    if(D->getParentFunctionOrMethod())
     {
         // skip function-local declarations
         return true;
     }
 
-    llvm::SmallString<128> filePath;
-    filePath = sm.getPresumedLoc(D->getBeginLoc()).getFilename();
-    if(config_.shouldSkipFile(filePath))
-        return true;
+    llvm::SmallString<512> filePath;
+    clang::PresumedLoc const loc =
+        sm.getPresumedLoc(D->getBeginLoc());
+    auto result = fileFilter_.emplace(
+        loc.getIncludeLoc().getRawEncoding(),
+        FileFilter());
+    if(result.second)
+    {
+        // new element
+        filePath = loc.getFilename();
+        FileFilter& ff = result.first->second;
+        ff.exclude = config_.filterFile(filePath, ff.prefix);
+        if(ff.exclude)
+            return true;
+        path::replace_path_prefix(filePath, ff.prefix, "");
+    }
+    else
+    {
+        // existing element
+        FileFilter const& ff = result.first->second;
+        if(ff.exclude)
+            return true;
+        filePath = loc.getFilename();
+        path::replace_path_prefix(filePath, ff.prefix, "");
+    }
 
     // If there is an error generating a USR for the decl, skip this decl.
     {
@@ -72,14 +95,13 @@ mapDecl(T const* D)
         }
     }
 
-    bool IsFileInRootDir;
-    llvm::SmallString<128> File =
-        getFile(D, D->getASTContext(), config_.SourceRoot, IsFileInRootDir);
+    // VFALCO is this right?
+    bool const IsFileInRootDir = true;
     auto I = emitInfo(
         D,
         getComment(D, D->getASTContext()),
         getLine(D, D->getASTContext()),
-        File,
+        filePath,
         IsFileInRootDir,
         config_.PublicOnly);
 
@@ -178,31 +200,5 @@ getLine(
         D->getBeginLoc()).getLine();
 }
 
-llvm::SmallString<128>
-Visitor::
-getFile(NamedDecl const* D,
-    ASTContext const& Context,
-    llvm::StringRef RootDir,
-    bool& IsFileInRootDir) const
-{
-    llvm::SmallString<128> File(
-        Context.getSourceManager()
-            .getPresumedLoc(D->getBeginLoc())
-            .getFilename());
-    IsFileInRootDir = false;
-    if (RootDir.empty() || !File.startswith(RootDir))
-        return File;
-    IsFileInRootDir = true;
-    llvm::SmallString<128> Prefix(RootDir);
-    // replace_path_prefix removes the exact prefix provided. The result of
-    // calling that function on ("A/B/C.c", "A/B", "") would be "/C.c", which
-    // starts with a / that is not needed. This is why we fix Prefix so it always
-    // ends with a / and the result has the desired format.
-    if (!llvm::sys::path::is_separator(Prefix.back()))
-        Prefix += llvm::sys::path::get_separator();
-    llvm::sys::path::replace_path_prefix(File, Prefix, "");
-    return File;
-}
-
 } // mrdox
 } // clang

source/lib/XML.cpp

@@ -352,9 +352,11 @@ writeInfo(
     Info const& I)
 {
     writeTagLine("extract-name", I.extractName());
+#if 0
     auto relPath = I.getRelativeFilePath(config_.SourceRoot);
     if(! relPath.empty())
         writeTagLine("rel-path", relPath);
+#endif
     writeTagLine("base-name", I.getFileBaseName());
 }
 

source/tests/TestMain.cpp

@@ -191,10 +191,19 @@ testResult(
 int
 testMain(int argc, const char** argv)
 {
+    namespace path = llvm::sys::path;
+
     Config config;
     Reporter R;
-    auto const gen = makeXMLGenerator();
+ 
+    for(int i = 1; i < argc; ++i)
+    {
+        llvm::SmallString<0> s(argv[i]);
+        path::remove_dots(s, true);
+        config.includePaths.emplace_back(std::move(s));
+    }
 
+    auto const gen = makeXMLGenerator();
     llvm::ThreadPool Pool(llvm::hardware_concurrency(
         tooling::ExecutorConcurrency));
     for(int i = 1; i < argc; ++i)

source/tool/ToolMain.cpp

@@ -69,15 +69,6 @@ OutDirectory(
     llvm::cl::init("."),
     llvm::cl::cat(ToolCategory));
 
-static
-llvm::cl::opt<std::string>
-SourceRoot(
-    "source-root", llvm::cl::desc(R"(
-Directory where processed files are stored.
-Links to definition locations will only be
-generated if the file is in this dir.)"),
-    llvm::cl::cat(ToolCategory));
-
 static
 llvm::cl::opt<std::string>
     ConfigPath(

testfiles/2.xml

@@ -4,11 +4,9 @@
 </all>
 <namespace name="">
     <extract-name>GlobalNamespace</extract-name>
-    <rel-path>GlobalNamespace</rel-path>
     <base-name>index</base-name>
     <namespace name="N" usr="rjlMNXAaWNMkQYMTJzRA1DR0DiE=">
         <extract-name>N</extract-name>
-        <rel-path>N</rel-path>
         <base-name>index</base-name>
     </namespace>
 </namespace>