Merge pull request #10 from kota7/develop

Develop, fixed typo to fix #9

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: striprtf
 Type: Package
 Title: Extract Text from RTF File
-Version: 0.4.5
+Version: 0.4.6
 Authors@R: c(
     person("Kota", "Mori", email = "[email protected]", role = c("aut", "cre"))
   )

NEWS

@@ -1,3 +1,10 @@
+striprtf v0.4.6 (Release date: 2017-12-04)
+==============
+
+* fixed bugs for parsing lower quote (`\u2018`)
+
+
+
 striprtf v0.4.5 (Release date: 2017-07-04)
 ==============
 

R/RcppExports.R

@@ -2,14 +2,14 @@
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
 to_hexstr <- function(x, pad = 4L) {
-    .Call('striprtf_to_hexstr', PACKAGE = 'striprtf', x, pad)
+    .Call('_striprtf_to_hexstr', PACKAGE = 'striprtf', x, pad)
 }
 
 hex_to_int <- function(h, sep = 'x') {
-    .Call('striprtf_hex_to_int', PACKAGE = 'striprtf', h, sep)
+    .Call('_striprtf_hex_to_int', PACKAGE = 'striprtf', h, sep)
 }
 
 strip_helper <- function(match_mat, dest_names, special_keys, special_hex, verbose) {
-    .Call('striprtf_strip_helper', PACKAGE = 'striprtf', match_mat, dest_names, special_keys, special_hex, verbose)
+    .Call('_striprtf_strip_helper', PACKAGE = 'striprtf', match_mat, dest_names, special_keys, special_hex, verbose)
 }
 

R/global.R

@@ -86,7 +86,7 @@
            # - \cell ... end of each cell
            "row", "cell"),
   hexstr = c("x2022", "x2014", "x2003", "x2013", "x2002",
-             "x201C", "x000A", "xu2018", "x000Ax000A", "x000A",
+             "x201C", "x000A", "x2018", "x000Ax000A", "x000A",
              "x2005", "x201D", "x2019", "x000Ax000A", "x0009",
              "x000A", "x0009"),
   str = c("\u2022", "\u2014", "\u2003", "\u2013", "\u2002",

README.md

@@ -1,4 +1,5 @@
 
+
 <!-- README.md is generated from README.Rmd. Please edit that file -->
 [![Build Status](https://travis-ci.org/kota7/striprtf.svg?branch=master)](https://travis-ci.org/kota7/striprtf) [![CRAN Status](https://www.r-pkg.org/badges/version/striprtf)](https://www.r-pkg.org/badges/version/striprtf) [![AppVeyor Build Status](https://ci.appveyor.com/api/projects/status/github/kota7/striprtf?branch=master&svg=true)](https://ci.appveyor.com/project/kota7/striprtf) [![](http://cranlogs.r-pkg.org/badges/striprtf)](https://cran.r-project.org/package=striprtf)
 

src/RcppExports.cpp

@@ -7,7 +7,7 @@ using namespace Rcpp;
 
 // to_hexstr
 std::string to_hexstr(int x, int pad);
-RcppExport SEXP striprtf_to_hexstr(SEXP xSEXP, SEXP padSEXP) {
+RcppExport SEXP _striprtf_to_hexstr(SEXP xSEXP, SEXP padSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
@@ -19,7 +19,7 @@ END_RCPP
 }
 // hex_to_int
 IntegerVector hex_to_int(std::string h, char sep);
-RcppExport SEXP striprtf_hex_to_int(SEXP hSEXP, SEXP sepSEXP) {
+RcppExport SEXP _striprtf_hex_to_int(SEXP hSEXP, SEXP sepSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
@@ -31,7 +31,7 @@ END_RCPP
 }
 // strip_helper
 List strip_helper(CharacterMatrix match_mat, CharacterVector dest_names, CharacterVector special_keys, CharacterVector special_hex, bool verbose);
-RcppExport SEXP striprtf_strip_helper(SEXP match_matSEXP, SEXP dest_namesSEXP, SEXP special_keysSEXP, SEXP special_hexSEXP, SEXP verboseSEXP) {
+RcppExport SEXP _striprtf_strip_helper(SEXP match_matSEXP, SEXP dest_namesSEXP, SEXP special_keysSEXP, SEXP special_hexSEXP, SEXP verboseSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
@@ -44,3 +44,15 @@ BEGIN_RCPP
     return rcpp_result_gen;
 END_RCPP
 }
+
+static const R_CallMethodDef CallEntries[] = {
+    {"_striprtf_to_hexstr", (DL_FUNC) &_striprtf_to_hexstr, 2},
+    {"_striprtf_hex_to_int", (DL_FUNC) &_striprtf_hex_to_int, 2},
+    {"_striprtf_strip_helper", (DL_FUNC) &_striprtf_strip_helper, 5},
+    {NULL, NULL, 0}
+};
+
+RcppExport void R_init_striprtf(DllInfo *dll) {
+    R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
+    R_useDynamicSymbols(dll, FALSE);
+}

src/dechex.cpp

@@ -40,21 +40,23 @@ IntegerVector hex_to_int(std::string h, char sep)
   //
   // returns:
   //   integer vector
+  //
+  // error if:
+  //   h contains a letter other than 0-9, a-f, A-F, and sep
 
   h += sep; // make sure the last hex is read
 
   IntegerVector out;
   bool started = false;
+  int start = 0;
   for (int i = 0; i < h.size(); i++)
   {
-    int start;
     if (h[i] == sep) {
       if (!started) {
         start = i + 1;
         started = true;
       } else {
         int end = i;
-
         int tmp = 0;
         int base = 1;
         for (int k = end-1; k >= start; k--)
@@ -64,8 +66,12 @@ IntegerVector hex_to_int(std::string h, char sep)
           if (c >= '0' && c <= '9')      n = c - '0';
           else if (c >= 'A' && c <= 'F') n = c - 'A' + 10;
           else if (c >= 'a' && c <= 'f') n = c - 'a' + 10;
-          else stop("invalid hex");
-
+          else {
+            std::string mess = "invalid hex found: ";
+            mess += c;
+            mess += ", around: " + h.substr(start, 10);
+            stop(mess);
+          }
           tmp += base*n;
           base *= 16;
         }

src/strip_helper.cpp

@@ -214,6 +214,7 @@ List strip_helper(CharacterMatrix match_mat,
     if (curhex.size() > 0 && (hex == "" || curhex.size() == 4)) {
       // make sure the length is 4
       while (curhex.size() < 4) curhex = '0' + curhex;
+      //Rcout << curhex << "\n";
       append_out(doc, 'x' + curhex, true, intable);
       curhex = "";
     }
@@ -249,6 +250,7 @@ List strip_helper(CharacterMatrix match_mat,
       } else if (ignorable) {
         continue;
       } else if (specialchars.haskey(word)) {
+        //Rcout << specialchars.getvalue(word) << "\n";
         append_out(doc, specialchars.getvalue(word), false, intable);
         if (word == "row") {
           intable = false;