hunspell: Fix Russian dictionarry in emacs.

7 years ago · 2d3744e119
1 changed files with 130 additions and 0 deletions
--- a/packages/hunspell/hunspell-chenc.patch
+++ b/packages/hunspell/hunspell-chenc.patch
@ -0,0 +1,130 @@
+Fix chenc and its usages
+chenc was changing buffer it was passed to. This caused untrackable
+multiple conversions of token and possibly other variables.
+
+Change it to only return converted string
+
+Additionally logic extending dest buffer implicitly assumed that 0
+bytes are left in dest buffer. It's not necessarily the case when
+converting to UTF-8 as if result would be 2-byte sequence and only 1 byte
+is remaining you get E2BIG as well.
+
+This fixes the case of pipe (-a) in UTF-8 with KOI8-R dictionary.
+
+diff -ur hunspell-1.6.1/src/tools/hunspell.cxx hunspell-1.6.1-mod2/src/tools/hunspell.cxx
+--- hunspell-1.6.1/src/tools/hunspell.cxx	2017-03-25 22:20:45.000000000 +0100
+++ hunspell-1.6.1-mod2/src/tools/hunspell.cxx	2017-08-14 23:22:16.246966174 +0200
+@@ -243,7 +243,7 @@
+ #endif
+ 
+ /* change character encoding */
+-std::string& chenc(std::string& st, const char* enc1, const char* enc2) {
+std::string chenc(const std::string& st, const char* enc1, const char* enc2) {
+ #ifndef HAVE_ICONV
+   (void)enc1;
+   (void)enc2;
+@@ -258,7 +258,7 @@
+   std::string out(st.size(), std::string::value_type());
+   size_t c1(st.size());
+   size_t c2(out.size());
+-  ICONV_CONST char* source = &st[0];
+  ICONV_CONST char* source = (ICONV_CONST char*) &st[0];
+   char* dest = &out[0];
+   iconv_t conv = iconv_open(fix_encoding_name(enc2), fix_encoding_name(enc1));
+   if (conv == (iconv_t)-1) {
+@@ -267,9 +267,10 @@
+     size_t res;
+     while ((res = iconv(conv, &source, &c1, &dest, &c2)) == size_t(-1)) {
+       if (errno == E2BIG) {
+        ssize_t destoff = dest - const_cast<char*>(&out[0]);
+         out.resize(out.size() + (c2 += c1));
+ 
+-        dest = const_cast<char*>(&out[0]) + out.size() - c2;
+        dest = const_cast<char*>(&out[0]) + destoff;
+       } else
+         break;
+     }
+@@ -278,7 +279,7 @@
+     }
+     iconv_close(conv);
+     out.resize(dest - &out[0]);
+-    st = out;
+    return out;
+   }
+ 
+   return st;
+@@ -507,8 +508,7 @@
+ #endif
+ 
+ int putdic(const std::string& in_word, Hunspell* pMS) {
+-  std::string word(in_word);
+-  chenc(word, ui_enc, dic_enc[0]);
+  std::string word = chenc(in_word, ui_enc, dic_enc[0]);
+ 
+   std::string buf;
+   pMS->input_conv(word.c_str(), buf);
+@@ -565,7 +565,7 @@
+   if (!dic)
+     return 0;
+   for (size_t i = 0; i < w.size(); ++i) {
+-    chenc(w[i], io_enc, ui_enc);
+    w[i] = chenc(w[i], io_enc, ui_enc);
+     fprintf(dic, "%s\n", w[i].c_str());
+   }
+   fclose(dic);
+@@ -595,8 +595,7 @@
+ // check words in the dictionaries (and set first checked dictionary)
+ bool check(Hunspell** pMS, int* d, const std::string& token, int* info, std::string* root) {
+   for (int i = 0; i < dmax; ++i) {
+-    std::string buf(token);
+-    chenc(buf, io_enc, dic_enc[*d]);
+    std::string buf = chenc(token, io_enc, dic_enc[*d]);
+     mystrrep(buf, ENTITY_APOS, "'");
+     if (checkapos && buf.find('\'') != std::string::npos)
+       return false;
+@@ -937,7 +936,7 @@
+                 fprintf(stdout, "%s", chenc(wlst[0], dic_enc[d], io_enc).c_str());
+               }
+               for (size_t j = 1; j < wlst.size(); ++j) {
+-                fprintf(stdout, ", %s", chenc(wlst[j], dic_enc[d], io_enc).c_str());
+                  fprintf(stdout, ", %s", chenc(wlst[j], dic_enc[d], io_enc).c_str());
+               }
+               fprintf(stdout, "\n");
+               fflush(stdout);
+@@ -1194,8 +1193,7 @@
+ }
+ 
+ std::string lower_first_char(const std::string& token, const char* ioenc, int langnum) {
+-  std::string utf8str(token);
+-  chenc(utf8str, ioenc, "UTF-8");
+  std::string utf8str = chenc(token, ioenc, "UTF-8");
+   std::vector<w_char> u;
+   u8_u16(u, utf8str);
+   if (!u.empty()) {
+@@ -1206,8 +1204,7 @@
+   }
+   std::string scratch;
+   u16_u8(scratch, u);
+-  chenc(scratch, "UTF-8", ioenc);
+-  return scratch;
+  return chenc(scratch, "UTF-8", ioenc);
+ }
+ 
+ // for terminal interface
+@@ -1532,13 +1529,13 @@
+       std::vector<std::string> wlst;
+       dialogscreen(parser, token, filename, info, wlst);  // preview
+       refresh();
+-      std::string buf(token);
+-      wlst = pMS[d]->suggest(mystrrep(chenc(buf, io_enc, dic_enc[d]), ENTITY_APOS, "'").c_str());
+      std::string dicbuf = chenc(token, io_enc, dic_enc[d]);
+      wlst = pMS[d]->suggest(mystrrep(dicbuf, ENTITY_APOS, "'").c_str());
+       if (wlst.empty()) {
+         dialogexit = dialog(parser, pMS[d], token, filename, wlst, info);
+       } else {
+         for (size_t j = 0; j < wlst.size(); ++j) {
+-          chenc(wlst[j], dic_enc[d], io_enc);
+          wlst[j] = chenc(wlst[j], dic_enc[d], io_enc);
+         }
+         dialogexit = dialog(parser, pMS[d], token, filename, wlst, info);
+       }