Initial modularization with both Clang and VC hacks

Note: gave up on VC about half way though.

1 files changed, 116 insertions, 0 deletions

diff --git a/libbutl/manifest-parser.mxx b/libbutl/manifest-parser.mxx
new file mode 100644
index 0000000..c21ccc1
--- /dev/null
+++ b/libbutl/manifest-parser.mxx
@@ -0,0 +1,116 @@
+// file      : libbutl/manifest-parser.mxx -*- C++ -*-
+// copyright : Copyright (c) 2014-2017 Code Synthesis Ltd
+// license   : MIT; see accompanying LICENSE file
+
+#ifndef __cpp_modules
+#pragma once
+#endif
+
+// C includes.
+
+#ifndef __cpp_lib_modules
+#include <string>
+#include <iosfwd>
+#include <cstdint>   // uint64_t
+#include <utility>   // pair
+#include <stdexcept> // runtime_error
+#endif
+
+// Other includes.
+
+#ifdef __cpp_modules
+export module butl.manifest_parser;
+#ifdef __cpp_lib_modules
+import std.core;
+import std.io;
+#endif
+import butl.char_scanner;
+#else
+#include <libbutl/char-scanner.mxx>
+#endif
+
+#include <libbutl/export.hxx>
+
+LIBBUTL_MODEXPORT namespace butl
+{
+  class LIBBUTL_SYMEXPORT manifest_parsing: public std::runtime_error
+  {
+  public:
+    manifest_parsing (const std::string& name,
+                      std::uint64_t line,
+                      std::uint64_t column,
+                      const std::string& description);
+
+    std::string name;
+    std::uint64_t line;
+    std::uint64_t column;
+    std::string description;
+  };
+
+  class manifest_name_value
+  {
+  public:
+    std::string name;
+    std::string value;
+
+    std::uint64_t name_line;
+    std::uint64_t name_column;
+
+    std::uint64_t value_line;
+    std::uint64_t value_column;
+
+    bool
+    empty () const {return name.empty () && value.empty ();}
+  };
+
+  class LIBBUTL_SYMEXPORT manifest_parser: protected butl::char_scanner
+  {
+  public:
+    manifest_parser (std::istream& is, const std::string& name)
+        : char_scanner (is), name_ (name) {}
+
+    const std::string&
+    name () const {return name_;}
+
+    // The first returned pair is special "start-of-manifest" with
+    // empty name and value being the format version: {"", "<ver>"}.
+    // After that we have a sequence of ordinary pairs which are
+    // the manifest. At the end of the manifest we have the special
+    // "end-of-manifest" pair with empty name and value: {"", ""}.
+    // After that we can either get another start-of-manifest pair
+    // (in which case the whole sequence repeats from the beginning)
+    // or we get another end-of-manifest pair which signals the end
+    // of stream (aka EOF). To put it another way, the parse sequence
+    // always has the following form:
+    //
+    // ({"", "<ver>"} {"<name>", "<value>"}* {"", ""})* {"", ""}
+    //
+    manifest_name_value
+    next ();
+
+    // Split the manifest value, optionally followed by ';' character and a
+    // comment into the value/comment pair. Note that ';' characters in the
+    // value must be escaped by the backslash.
+    //
+    static std::pair<std::string, std::string>
+    split_comment (const std::string&);
+
+  private:
+    void
+    parse_name (manifest_name_value&);
+
+    void
+    parse_value (manifest_name_value&);
+
+    // Skip spaces and return the first peeked non-space character.
+    //
+    xchar
+    skip_spaces ();
+
+  private:
+    const std::string name_;
+
+    enum {start, body, end} s_ = start;
+    std::string version_; // Current format version.
+  };
+}