Merge pull request #5945 from afishhh/master

Make `nix search` highlight all regexes and matches

5 files changed, 154 insertions, 26 deletions

diff --git a/src/libutil/fmt.cc b/src/libutil/fmt.cc
new file mode 100644
index 000000000..914fb62b2
--- /dev/null
+++ b/src/libutil/fmt.cc
@@ -0,0 +1,38 @@
+#include <regex>
+
+namespace nix {
+
+std::string hiliteMatches(const std::string &s, std::vector<std::smatch> matches, std::string prefix, std::string postfix) {
+    // Avoid copy on zero matches
+    if (matches.size() == 0)
+        return s;
+
+    std::sort(matches.begin(), matches.end(), [](const auto &a, const auto &b) {
+        return a.position() < b.position();
+    });
+
+    std::string out;
+    ssize_t last_end = 0;
+
+    for (auto it = matches.begin(); it != matches.end();) {
+        auto m = *it;
+        size_t start = m.position();
+        out.append(s.substr(last_end, m.position() - last_end));
+        // Merge continous matches
+        ssize_t end = start + m.length();
+        while(++it != matches.end() && (*it).position() <= end) {
+            auto n = *it;
+            ssize_t nend = start + (n.position() - start + n.length());
+            if(nend > end)
+                end = nend;
+        }
+        out.append(prefix);
+        out.append(s.substr(start, end - start));
+        out.append(postfix);
+        last_end = end;
+    }
+    out.append(s.substr(last_end));
+    return out;
+}
+
+}
diff --git a/src/libutil/fmt.hh b/src/libutil/fmt.hh
index fd335b811..1f81bfcfb 100644
--- a/src/libutil/fmt.hh
+++ b/src/libutil/fmt.hh
@@ -2,6 +2,7 @@
 
 #include <boost/format.hpp>
 #include <string>
+#include <regex>
 #include "ansicolor.hh"
 
 
@@ -154,4 +155,13 @@ inline hintformat hintfmt(std::string plain_string)
     // we won't be receiving any args in this case, so just print the original string
     return hintfmt("%s", normaltxt(plain_string));
 }
+
+/**
+ * Highlight all the given matches in the given string `s` by wrapping them
+ * between `prefix` and `postfix`.
+ *
+ * If some matches overlap, then their union will be wrapped rather than the
+ * individual matches.
+ */
+std::string hiliteMatches(const std::string &s, std::vector<std::smatch> matches, std::string prefix, std::string postfix);
 }
diff --git a/src/libutil/tests/fmt.cc b/src/libutil/tests/fmt.cc
new file mode 100644
index 000000000..33772162c
--- /dev/null
+++ b/src/libutil/tests/fmt.cc
@@ -0,0 +1,68 @@
+#include "fmt.hh"
+
+#include <gtest/gtest.h>
+
+#include <regex>
+
+namespace nix {
+/* ----------- tests for fmt.hh -------------------------------------------------*/
+
+    TEST(hiliteMatches, noHighlight) {
+        ASSERT_STREQ(hiliteMatches("Hello, world!", std::vector<std::smatch>(), "(", ")").c_str(), "Hello, world!");
+    }
+
+    TEST(hiliteMatches, simpleHighlight) {
+        std::string str = "Hello, world!";
+        std::regex re = std::regex("world");
+        auto matches = std::vector(std::sregex_iterator(str.begin(), str.end(), re), std::sregex_iterator());
+        ASSERT_STREQ(
+                    hiliteMatches(str, matches, "(", ")").c_str(),
+                    "Hello, (world)!"
+        );
+    }
+
+    TEST(hiliteMatches, multipleMatches) {
+        std::string str = "Hello, world, world, world, world, world, world, Hello!";
+        std::regex re = std::regex("world");
+        auto matches = std::vector(std::sregex_iterator(str.begin(), str.end(), re), std::sregex_iterator());
+        ASSERT_STREQ(
+                    hiliteMatches(str, matches, "(", ")").c_str(),
+                    "Hello, (world), (world), (world), (world), (world), (world), Hello!"
+        );
+    }
+
+    TEST(hiliteMatches, overlappingMatches) {
+        std::string str = "world, Hello, world, Hello, world, Hello, world, Hello, world!";
+        std::regex re = std::regex("Hello, world");
+        std::regex re2 = std::regex("world, Hello");
+        auto v = std::vector(std::sregex_iterator(str.begin(), str.end(), re), std::sregex_iterator());
+        for(auto it = std::sregex_iterator(str.begin(), str.end(), re2); it != std::sregex_iterator(); ++it) {
+            v.push_back(*it);
+        }
+        ASSERT_STREQ(
+                    hiliteMatches(str, v, "(", ")").c_str(),
+                    "(world, Hello, world, Hello, world, Hello, world, Hello, world)!"
+        );
+    }
+
+    TEST(hiliteMatches, complexOverlappingMatches) {
+        std::string str = "legacyPackages.x86_64-linux.git-crypt";
+        std::vector regexes = {
+            std::regex("t-cry"),
+            std::regex("ux\\.git-cry"),
+            std::regex("git-c"),
+            std::regex("pt"),
+        };
+        std::vector<std::smatch> matches;
+        for(auto regex : regexes)
+        {
+            for(auto it = std::sregex_iterator(str.begin(), str.end(), regex); it != std::sregex_iterator(); ++it) {
+                matches.push_back(*it);
+            }
+        }
+        ASSERT_STREQ(
+                    hiliteMatches(str, matches, "(", ")").c_str(),
+                    "legacyPackages.x86_64-lin(ux.git-crypt)"
+        );
+    }
+}
diff --git a/src/nix/search.cc b/src/nix/search.cc
index 0d8fdd5c2..0d10d8c2e 100644
--- a/src/nix/search.cc
+++ b/src/nix/search.cc
@@ -9,6 +9,7 @@
 #include "shared.hh"
 #include "eval-cache.hh"
 #include "attr-path.hh"
+#include "fmt.hh"
 
 #include <regex>
 #include <fstream>
@@ -20,16 +21,6 @@ std::string wrap(std::string prefix, std::string s)
     return prefix + s + ANSI_NORMAL;
 }
 
-std::string hilite(const std::string & s, const std::smatch & m, std::string postfix)
-{
-    return
-        m.empty()
-        ? s
-        : std::string(m.prefix())
-          + ANSI_GREEN + std::string(m.str()) + postfix
-          + std::string(m.suffix());
-}
-
 struct CmdSearch : InstallableCommand, MixJSON
 {
     std::vector<std::string> res;
@@ -100,8 +91,6 @@ struct CmdSearch : InstallableCommand, MixJSON
                 };
 
                 if (cursor.isDerivation()) {
-                    size_t found = 0;
-
                     DrvName name(cursor.getAttr("name")->getString());
 
                     auto aMeta = cursor.maybeGetAttr("meta");
@@ -110,21 +99,31 @@ struct CmdSearch : InstallableCommand, MixJSON
                     std::replace(description.begin(), description.end(), '\n', ' ');
                     auto attrPath2 = concatStringsSep(".", attrPath);
 
-                    std::smatch attrPathMatch;
-                    std::smatch descriptionMatch;
-                    std::smatch nameMatch;
+                    std::vector<std::smatch> attrPathMatches;
+                    std::vector<std::smatch> descriptionMatches;
+                    std::vector<std::smatch> nameMatches;
+                    bool found = false;
 
                     for (auto & regex : regexes) {
-                        std::regex_search(attrPath2, attrPathMatch, regex);
-                        std::regex_search(name.name, nameMatch, regex);
-                        std::regex_search(description, descriptionMatch, regex);
-                        if (!attrPathMatch.empty()
-                            || !nameMatch.empty()
-                            || !descriptionMatch.empty())
-                            found++;
+                        found = false;
+                        auto add_all = [&found](std::sregex_iterator it, std::vector<std::smatch>& vec){
+                            const auto end = std::sregex_iterator();
+                            while(it != end) {
+                                vec.push_back(*it++);
+                                found = true;
+                            }
+                        };
+
+                        add_all(std::sregex_iterator(attrPath2.begin(), attrPath2.end(), regex), attrPathMatches);
+                        add_all(std::sregex_iterator(name.name.begin(), name.name.end(), regex), nameMatches);
+                        add_all(std::sregex_iterator(description.begin(), description.end(), regex), descriptionMatches);
+
+                        if(!found)
+                            break;
                     }
 
-                    if (found == res.size()) {
+                    if (found)
+                    {
                         results++;
                         if (json) {
                             auto jsonElem = jsonOut->object(attrPath2);
@@ -132,15 +131,15 @@ struct CmdSearch : InstallableCommand, MixJSON
                             jsonElem.attr("version", name.version);
                             jsonElem.attr("description", description);
                         } else {
-                            auto name2 = hilite(name.name, nameMatch, "\e[0;2m");
+                            auto name2 = hiliteMatches(name.name, std::move(nameMatches), ANSI_GREEN, "\e[0;2m");
                             if (results > 1) logger->cout("");
                             logger->cout(
                                 "* %s%s",
-                                wrap("\e[0;1m", hilite(attrPath2, attrPathMatch, "\e[0;1m")),
+                                wrap("\e[0;1m", hiliteMatches(attrPath2, std::move(attrPathMatches), ANSI_GREEN, "\e[0;1m")),
                                 name.version != "" ? " (" + name.version + ")" : "");
                             if (description != "")
                                 logger->cout(
-                                    "  %s", hilite(description, descriptionMatch, ANSI_NORMAL));
+                                    "  %s", hiliteMatches(description, std::move(descriptionMatches), ANSI_GREEN, ANSI_NORMAL));
                         }
                     }
                 }
diff --git a/tests/search.sh b/tests/search.sh
index ee3261687..52e12f381 100644
--- a/tests/search.sh
+++ b/tests/search.sh
@@ -23,3 +23,16 @@ clearCache
 nix search -f search.nix '' |grep -q foo
 nix search -f search.nix '' |grep -q bar
 nix search -f search.nix '' |grep -q hello
+
+## Tests for multiple regex/match highlighting
+
+e=$'\x1b' # grep doesn't support \e, \033 or even \x1b
+# Multiple overlapping regexes
+(( $(nix search -f search.nix '' 'oo' 'foo' 'oo' | grep "$e\[32;1mfoo$e\\[0;1m" | wc -l) == 1 ))
+(( $(nix search -f search.nix '' 'broken b' 'en bar' | grep "$e\[32;1mbroken bar$e\\[0m" | wc -l) == 1 ))
+
+# Multiple matches
+# Searching for 'o' should yield the 'o' in 'broken bar', the 'oo' in foo and 'o' in hello
+(( $(nix search -f search.nix '' 'o' | grep -Eo "$e\[32;1mo{1,2}$e\[(0|0;1)m" | wc -l) == 3 ))
+# Searching for 'b' should yield the 'b' in bar and the two 'b's in 'broken bar'
+(( $(nix search -f search.nix '' 'b' | grep -Eo "$e\[32;1mb$e\[(0|0;1)m" | wc -l) == 3 ))