From 235408636362a8c21752a869a1f603315b0d9e19 Mon Sep 17 00:00:00 2001
From: Karen Arutyunov <karen@codesynthesis.com>
Date: Sat, 22 Feb 2020 18:02:37 +0300
Subject: UTF-8-sanitize command logs

---
 bbot/worker/worker.cxx | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

diff --git a/bbot/worker/worker.cxx b/bbot/worker/worker.cxx
index f25a287..eb2609d 100644
--- a/bbot/worker/worker.cxx
+++ b/bbot/worker/worker.cxx
@@ -16,6 +16,7 @@
 
 #include <libbutl/b.mxx>
 #include <libbutl/pager.mxx>
+#include <libbutl/utility.mxx>       // to_utf8()
 #include <libbutl/filesystem.mxx>
 #include <libbutl/string-parser.mxx>
 
@@ -138,14 +139,11 @@ run_cmd (tracer& t,
 
     result_status r (result_status::success);
 
-    // Log the diagnostics. Also print it to stderr at verbosity level 3 or
-    // higher.
+    // UTF-8-sanitize and log the diagnostics. Also print the raw diagnostics
+    // to stderr at verbosity level 3 or higher.
     //
-    auto add = [&log, &t] (const string& s, bool trace = true)
+    auto add = [&log, &t] (string&& s, bool trace = true)
     {
-      log += s;
-      log += '\n';
-
       if (verb >= 3)
       {
         if (trace)
@@ -153,6 +151,11 @@ run_cmd (tracer& t,
         else
           text << s;
       }
+
+      to_utf8 (s, '?');
+
+      log += s;
+      log += '\n';
     };
 
     {
@@ -161,7 +164,6 @@ run_cmd (tracer& t,
       for (string l; is.peek () != ifdstream::traits_type::eof (); )
       {
         getline (is, l);
-        add (l, false);
 
         // Match the log line with the warning-detecting regular expressions
         // until the first match.
@@ -183,6 +185,8 @@ run_cmd (tracer& t,
             }
           }
         }
+
+        add (move (l), false /* trace */);
       }
     }
 
-- 
cgit v1.1