Fixed build on Linux and macOS that was failing because of using std::to_chars and std::from_chars with floating-point values.

The old version of GCC and Clang support only integers to be passed to std::to_chars and std::from_chars. macOS older version of Clang doesn't support std::from_chars at all. So for Linux and macOS, it was replaced std::from_chars with strtod and temporarily was replace std::to_chars with snprintf.

src/libslic3r/GCodeReader.cpp

@@ -71,9 +71,16 @@ const char* GCodeReader::parse_line_internal(const char *ptr, const char *end, G
             }
             if (axis != NUM_AXES_WITH_UNKNOWN) {
                 // Try to parse the numeric value.
+#ifdef WIN32
                 double v;
                 auto [pend, ec] = std::from_chars(++ c, end, v);
                 if (pend != c && is_end_of_word(*pend)) {
+#else
+                // The older version of GCC and Clang support std::from_chars just for integers, so strtod we used it instead.
+                char   *pend = nullptr;
+                double  v = strtod(++ c, &pend);
+                if (pend != nullptr && is_end_of_word(*pend)) {
+#endif
                     // The axis value has been parsed correctly.
                     if (axis != UNKNOWN_AXIS)
 	                    gline.m_axis[int(axis)] = float(v);

src/libslic3r/GCodeWriter.cpp

@@ -309,7 +309,16 @@ public:
 
     void emit_axis(const char axis, const double v, size_t digits) {
         *ptr_err.ptr ++ = ' '; *ptr_err.ptr ++ = axis;
+#ifdef WIN32
         this->ptr_err = std::to_chars(this->ptr_err.ptr, this->buf_end, v, std::chars_format::fixed, digits);
+#else
+        int buf_capacity = int(this->buf_end - this->ptr_err.ptr);
+        int ret          = snprintf(this->ptr_err.ptr, buf_capacity, "%.*lf", int(digits), v);
+        if (ret <= 0 || ret > buf_capacity)
+            ptr_err.ec = std::errc::value_too_large;
+        else
+            this->ptr_err.ptr = this->ptr_err.ptr + ret;
+#endif
     }
 
     void emit_xy(const Vec2d &point) {