From 3f8ef886b1e2b17d8ba12a74355ab7e09aec5ed6 Mon Sep 17 00:00:00 2001
From: Karen Arutyunov <karen@codesynthesis.com>
Date: Fri, 6 Mar 2020 17:40:45 +0300
Subject: Restrict buildfiles and testscripts with only Unicode graphic
 characters, '\t', '\r', and '\n'

---
 libbuild2/lexer.hxx | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/libbuild2/lexer.hxx b/libbuild2/lexer.hxx
index 93992b4..e9a3149 100644
--- a/libbuild2/lexer.hxx
+++ b/libbuild2/lexer.hxx
@@ -7,6 +7,7 @@
 #include <stack>
 
 #include <libbutl/utf8.mxx>
+#include <libbutl/unicode.mxx>
 #include <libbutl/char-scanner.mxx>
 
 #include <libbuild2/types.hxx>
@@ -201,10 +202,14 @@ namespace build2
     lexer (istream& is, const path_name& name, uint64_t line,
            const char* escapes,
            bool set_mode)
-        : char_scanner (is, true /* crlf */, line),
-          fail ("error", &name),
-          name_ (name),
-          sep_ (false)
+      : char_scanner (is,
+                      butl::utf8_validator (butl::codepoint_types::graphic,
+                                            U"\n\r\t"),
+                      true /* crlf */,
+                      line),
+        fail ("error", &name),
+        name_ (name),
+        sep_ (false)
     {
       if (set_mode)
         mode (lexer_mode::normal, '@', escapes);
-- 
cgit v1.1