From f063bdfd8ef8d27a9bc0651aa19da5fb2c10eb76 Mon Sep 17 00:00:00 2001
From: Danny Robson <danny@nerdcruft.net>
Date: Fri, 26 Apr 2019 09:08:03 +1000
Subject: [PATCH] tools/compare: allow multiple sources

---
 tools/compare.cpp | 102 ++++++++++++++++++++++++++++++++++------------
 1 file changed, 76 insertions(+), 26 deletions(-)

diff --git a/tools/compare.cpp b/tools/compare.cpp
index 8d0c689..66a2c44 100644
--- a/tools/compare.cpp
+++ b/tools/compare.cpp
@@ -39,7 +39,7 @@ namespace emory {
         };
 
 
-        struct digest_comparator {
+        struct digest_ordering {
             bool operator() (chunk const &a, chunk const &b)
             {
                 return a.digest < b.digest;
@@ -48,6 +48,30 @@ namespace emory {
     };
 
 
+    template <typename IdT>
+    struct match {
+        struct {
+            IdT id;
+            chunk data;
+        } src, dst;
+
+        struct src_equality {
+            bool operator() (match const &a, match const &b)
+            {
+                return a.src == b.src;
+            }
+        };
+
+
+        struct src_ordering {
+            bool operator() (match const &a, match const &b)
+            {
+                return a.src < b.src;
+            }
+        };
+    };
+
+
     std::ostream& operator<< (std::ostream &os, chunk const &val)
     {
         return os << "{ first: 0x"  << std::setw (8) << std::setfill ('0') << std::hex << val.offset.first
@@ -62,7 +86,7 @@ namespace emory {
     };
 
 
-    std::vector<chunk> common (map const &a, map const &b);
+    std::vector<match<map const*>> common (map const &a, map const &b);
 }
 
 
@@ -100,31 +124,36 @@ emory::map::map (cruft::view<u08 const *> data, const emory::params &p)
 }
 
 
-std::vector<emory::chunk>
-emory::common (emory::map const &a, emory::map const &b)
+std::vector<
+    emory::match<emory::map const*>
+>
+emory::common (emory::map const &src, emory::map const &dst)
 {
-    CHECK (std::is_sorted (a.elements.begin (), a.elements.end (), emory::chunk::digest_comparator {}));
-    CHECK (std::is_sorted (b.elements.begin (), b.elements.end (), emory::chunk::digest_comparator {}));
+    CHECK (std::is_sorted (src.elements.begin (), src.elements.end (), emory::chunk::digest_ordering {}));
+    CHECK (std::is_sorted (dst.elements.begin (), dst.elements.end (), emory::chunk::digest_ordering {}));
 
-    std::vector<emory::chunk> res;
+    std::vector<emory::match<emory::map const*>> res;
 
-    for (auto a_cursor = a.elements.begin (), b_cursor = b.elements.begin ();
-         a_cursor != a.elements.end () && b_cursor != b.elements.end ();
+    for (auto src_cursor = src.elements.begin (), dst_cursor = dst.elements.begin ();
+         src_cursor != src.elements.end () && dst_cursor != dst.elements.end ();
          /* nothing */)
     {
-        if (a_cursor->digest < b_cursor->digest) {
-            ++a_cursor;
+        if (src_cursor->digest < dst_cursor->digest) {
+            ++src_cursor;
             continue;
         }
 
-        if (b_cursor->digest < a_cursor->digest) {
-            ++b_cursor;
+        if (dst_cursor->digest < src_cursor->digest) {
+            ++dst_cursor;
             continue;
         }
 
-        res.push_back (*a_cursor);
-        ++a_cursor;
-        ++b_cursor;
+        res.push_back ({
+            .src = { .id = &src, .data = *src_cursor },
+            .dst = { .id = &dst, .data = *dst_cursor },
+        });
+        ++src_cursor;
+        ++dst_cursor;
     }
 
     return res;
@@ -145,8 +174,8 @@ enum {
 
 int main (int argc, char const **argv)
 {
-    if (argc != NUM_ARGS) {
-        std::cerr << "usage: " << argv[ARG_SELF] << " <bits> <window> <minimum> <target> <source>\n";
+    if (argc < NUM_ARGS) {
+        std::cerr << "usage: " << argv[ARG_SELF] << " <bits> <window> <minimum> <target> <source> [...]\n";
         return EXIT_FAILURE;
     }
 
@@ -156,16 +185,37 @@ int main (int argc, char const **argv)
         .minimum = cruft::parse::from_string<std::ptrdiff_t> (argv[ARGS_MINIMUM]),
     };
 
-    emory::map target (cruft::mapped_file (argv[ARGS_TARGET]), p);
-    emory::map source (cruft::mapped_file (argv[ARGS_SOURCE]), p);
-
     std::clog << "Hashing target\n";
-    std::sort (target.elements.begin (), target.elements.end (), emory::chunk::digest_comparator {});
-    std::clog << "Hashing source\n";
-    std::sort (source.elements.begin (), source.elements.end (), emory::chunk::digest_comparator {});
+    emory::map target (cruft::mapped_file (argv[ARGS_TARGET]), p);
+    std::sort (target.elements.begin (), target.elements.end (), emory::chunk::digest_ordering {});
 
-    std::clog << "Finding common\n";
-    auto const &found = common (target, source);
+    std::vector<emory::chunk> found;
+
+    for (int i = ARGS_SOURCE; i != argc; ++i) {
+        auto const path = argv[i];
+        std::clog << "Hashing source: " << path << '\n';
+        emory::map source (cruft::mapped_file (path), p);
+        std::sort (source.elements.begin (), source.elements.end (), emory::chunk::digest_ordering {});
+
+        std::clog << "Finding common\n";
+        auto const &source_matches = common (target, source);
+        std::clog << "Discovered " << source_matches.size () << " blocks\n";
+
+        std::transform (
+            std::begin (source_matches),
+            std::end   (source_matches),
+            std::back_inserter (found),
+            [] (auto const &i) { return i.src.data; }
+        );
+
+        std::sort (std::begin (found), std::end (found), emory::chunk::digest_ordering {});
+        found.erase (
+            std::unique (std::begin (found), std::end (found), emory::chunk::digest_equality {}),
+            found.end ()
+        );
+    }
+
+    std::clog << "Finalising\n";
     std::size_t matching = 0;
     for (auto const &i: found) {
         std::cout << i << '\n';