Improve filtering of ANSI escape sequences in build logs

All ANSI sequences except color setting are now filtered out. In
particular, terminal resets (such as from NixOS VM tests) are filtered
out.

Also, fix the completely broken tab character handling.

src/libstore/build.cc

@@ -3428,7 +3428,7 @@ void DerivationGoal::flushLine()
     else {
         if (settings.verboseBuild &&
             (settings.printRepeatedBuilds || curRound == 1))
-            printError(filterANSIEscapes(currentLogLine, true));
+            printError(currentLogLine);
         else {
             logTail.push_back(currentLogLine);
             if (logTail.size() > settings.logLines) logTail.pop_front();

src/libutil/logging.cc

@@ -44,7 +44,7 @@ public:
             prefix = std::string("<") + c + ">";
         }
 
-        writeToStderr(prefix + (tty ? fs.s : filterANSIEscapes(fs.s)) + "\n");
+        writeToStderr(prefix + filterANSIEscapes(fs.s) + "\n");
     }
 
     void startActivity(ActivityId act, Verbosity lvl, ActivityType type,

src/libutil/util.cc

@@ -1178,36 +1178,51 @@ void ignoreException()
 }
 
 
-string filterANSIEscapes(const string & s, bool nixOnly)
+std::string filterANSIEscapes(const std::string & s, unsigned int width)
 {
-    string t, r;
-    enum { stTop, stEscape, stCSI } state = stTop;
-    for (auto c : s) {
-        if (state == stTop) {
-            if (c == '\e') {
-                state = stEscape;
-                r = c;
-            } else
-                t += c;
-        } else if (state == stEscape) {
-            r += c;
-            if (c == '[')
-                state = stCSI;
-            else {
-                t += r;
-                state = stTop;
+    std::string t, e;
+    size_t w = 0;
+    auto i = s.begin();
+
+    while (w < (size_t) width && i != s.end()) {
+
+        if (*i == '\e') {
+            std::string e;
+            e += *i++;
+            char last = 0;
+
+            if (i != s.end() && *i == '[') {
+                e += *i++;
+                // eat parameter bytes
+                while (i != s.end() && *i >= 0x30 && *i <= 0x3f) e += *i++;
+                // eat intermediate bytes
+                while (i != s.end() && *i >= 0x20 && *i <= 0x2f) e += *i++;
+                // eat final byte
+                if (i != s.end() && *i >= 0x40 && *i <= 0x7e) e += last = *i++;
+            } else {
+                if (i != s.end() && *i >= 0x40 && *i <= 0x5f) e += *i++;
             }
-        } else {
-            r += c;
-            if (c >= 0x40 && c <= 0x7e) {
-                if (nixOnly && (c != 'p' && c != 'q' && c != 's' && c != 'a' && c != 'b'))
-                    t += r;
-                state = stTop;
-                r.clear();
+
+            if (last == 'm')
+                t += e;
+        }
+
+        else if (*i == '\t') {
+            i++; t += ' '; w++;
+            while (w < (size_t) width && w % 8) {
+                t += ' '; w++;
             }
         }
+
+        else if (*i == '\r')
+            // do nothing for now
+            ;
+
+        else {
+            t += *i++; w++;
+        }
     }
-    t += r;
+
     return t;
 }
 

src/libutil/util.hh

@@ -388,10 +388,12 @@ void ignoreException();
 #define ANSI_BLUE "\e[34;1m"
 
 
-/* Filter out ANSI escape codes from the given string. If ‘nixOnly’ is
-   set, only filter escape codes generated by Nixpkgs' stdenv (used to
-   denote nesting etc.). */
-string filterANSIEscapes(const string & s, bool nixOnly = false);
+/* Truncate a string to 'width' printable characters. Certain ANSI
+   escape sequences (such as colour setting) are copied but not
+   included in the character count. Other ANSI escape sequences are
+   filtered. Also, tabs are expanded to spaces. */
+std::string filterANSIEscapes(const std::string & s,
+    unsigned int width = std::numeric_limits<unsigned int>::max());
 
 
 /* Base64 encoding/decoding. */

src/nix/progress-bar.cc

@@ -23,44 +23,6 @@ static uint64_t getI(const std::vector<Logger::Field> & fields, size_t n)
     return fields[n].i;
 }
 
-/* Truncate a string to 'width' printable characters. ANSI escape
-   sequences are copied but not included in the character count. Also,
-   tabs are expanded to spaces. */
-static std::string ansiTruncate(const std::string & s, int width)
-{
-    if (width <= 0) return s;
-
-    std::string t;
-    size_t w = 0;
-    auto i = s.begin();
-
-    while (w < (size_t) width && i != s.end()) {
-        if (*i == '\e') {
-            t += *i++;
-            if (i != s.end() && *i == '[') {
-                t += *i++;
-                while (i != s.end() && (*i < 0x40 || *i > 0x7e)) {
-                    t += *i++;
-                }
-                if (i != s.end()) t += *i++;
-            }
-        }
-
-        else if (*i == '\t') {
-            t += ' '; w++;
-            while (w < (size_t) width && w & 8) {
-                t += ' '; w++;
-            }
-        }
-
-        else {
-            t += *i++; w++;
-        }
-    }
-
-    return t;
-}
-
 class ProgressBar : public Logger
 {
 private:
@@ -343,7 +305,10 @@ public:
             }
         }
 
-        writeToStderr("\r" + ansiTruncate(line, getWindowSize().second) + "\e[K");
+        auto width = getWindowSize().second;
+        if (width <= 0) std::numeric_limits<decltype(width)>::max();
+
+        writeToStderr("\r" + filterANSIEscapes(line, width) + "\e[K");
     }
 
     std::string getStatus(State & state)

tests/misc.sh

@@ -16,4 +16,4 @@ nix-env --foo 2>&1 | grep "no operation"
 nix-env -q --foo 2>&1 | grep "unknown flag"
 
 # Eval Errors.
-nix-instantiate --eval -E 'let a = {} // a; in a.foo' 2>&1 | grep "infinite recursion encountered, at (string):1:15$"
+nix-instantiate --eval -E 'let a = {} // a; in a.foo' 2>&1 | grep "infinite recursion encountered, at .*(string).*:1:15$"