From a11916449656b3d514930d49827c34e5d499729f Mon Sep 17 00:00:00 2001
From: Axel Kohlmeyer <akohlmey@gmail.com>
Date: Mon, 25 Jan 2021 20:53:59 -0500
Subject: [PATCH] add utf-8 check, substitution, and warning to Input::parse()
 function

---
 src/input.cpp | 11 +++++++++++
 src/input.h   |  1 +
 2 files changed, 12 insertions(+)

diff --git a/src/input.cpp b/src/input.cpp
index 457cf74b1a..df5cf0efbe 100644
--- a/src/input.cpp
+++ b/src/input.cpp
@@ -106,6 +106,7 @@ Input::Input(LAMMPS *lmp, int argc, char **argv) : Pointers(lmp)
   label_active = 0;
   labelstr = nullptr;
   jump_skip = 0;
+  utf8_warn = true;
 
   if (me == 0) {
     nfile = 1;
@@ -421,6 +422,16 @@ void Input::parse()
     ptr++;
   }
 
+  if (utils::has_utf8(copy)) {
+    std::string buf = utils::utf8_subst(copy);
+    strcpy(copy,buf.c_str());
+    if (utf8_warn && (comm->me == 0))
+      error->warning(FLERR,"Detected non-ASCII characters in input. "
+                     "Will try to continue by replacing with ASCII "
+                     "equivalents where known.");
+    utf8_warn = false;
+  }
+
   // perform $ variable substitution (print changes)
   // except if searching for a label since earlier variable may not be defined
 
diff --git a/src/input.h b/src/input.h
index b8ffb276f9..a86b10a686 100644
--- a/src/input.h
+++ b/src/input.h
@@ -54,6 +54,7 @@ class Input : protected Pointers {
   int label_active;            // 0 = no label, 1 = looking for label
   char *labelstr;              // label string being looked for
   int jump_skip;               // 1 if skipping next jump, 0 otherwise
+  bool utf8_warn;              // true if need to warn about UTF-8 chars
 
   FILE **infiles;              // list of open input files