by default use codepage1252 for metadata including non-ascii characters while missing codepage info

* IMCtermite.pyx: add codepage conversion for windows
* bump VERSION

.github/workflows/pypi-deploy.yml

@@ -40,15 +40,6 @@ jobs:
         with:
           name: wheel-config
           path: python/
-      - name: Get vcpkg repo
-        if: matrix.os == 'windows-2019'
-        run: git clone https://github.com/Microsoft/vcpkg.git
-      - name: Build and install vcpkg and install libiconv for windows
-        if: matrix.os == 'windows-2019'
-        run: |
-          .\vcpkg\bootstrap-vcpkg.bat
-          vcpkg install libiconv
-        working-directory: vcpkg/
       - name: Build wheels
         run: python -m cibuildwheel --output-dir wheelhouse
         working-directory: python/

lib/imc_channel.hpp

@@ -10,7 +10,11 @@
 #include <chrono>
 #include <ctime>
 #include <time.h>
+#if defined(__linux__) || defined(__APPLE__)
 #include <iconv.h>
+#elif defined(__WIN32__) || defined(_WIN32)
+#define timegm _mkgmtime
+#endif
 
 //---------------------------------------------------------------------------//
 
@@ -144,12 +148,13 @@ namespace imc
     return sumstr;
   }
 
+  #if defined(__linux__) || defined(__APPLE__)
   // convert encoding of any descriptions, channel-names, units etc.
   class iconverter
   {
-      std::string in_enc_, out_enc_;
-      iconv_t cd_;
-      size_t out_buffer_size_;
+    std::string in_enc_, out_enc_;
+    iconv_t cd_;
+    size_t out_buffer_size_;
 
     public:
 
@@ -215,6 +220,14 @@ namespace imc
         astring = outstring;
       }
   };
+  #elif defined(__WIN32__) || defined(_WIN32)
+  class iconverter
+  {
+    public:
+      iconverter(std::string in_enc, std::string out_enc, size_t out_buffer_size = 1024) {}
+      void convert(std::string &astring) {}
+  };
+  #endif
 
   // channel
   struct channel
@@ -483,28 +496,35 @@ namespace imc
     // convert any description, units etc. to UTF-8 (by default)
     void convert_encoding()
     {
+      // actual input codepage
+      std::string cpn;
+
       if ( !codepage_.empty() )
       {
         // construct iconv-compatible name for respective codepage
-        std::string cpn = std::string("CP") + codepage_;
-
-        // set up converter
-        std::string utf = std::string("UTF-8");
-        iconverter conv(cpn,utf);
-
-        conv.convert(name_);
-        conv.convert(comment_);
-        conv.convert(origin_);
-        conv.convert(origin_comment_);
-        conv.convert(text_);
-        conv.convert(language_code_);
-        conv.convert(yname_);
-        conv.convert(yunit_);
-        conv.convert(xname_);
-        conv.convert(xunit_);
-        conv.convert(group_name_);
-        conv.convert(group_comment_);
+        cpn = std::string("CP") + codepage_;
       }
+      else {
+	// assume codepage 1252 by default
+	cpn = std::string("CP1252");
+      }
+
+      // set up converter
+      std::string utf = std::string("UTF-8");
+      iconverter conv(cpn,utf);
+
+      conv.convert(name_);
+      conv.convert(comment_);
+      conv.convert(origin_);
+      conv.convert(origin_comment_);
+      conv.convert(text_);
+      conv.convert(language_code_);
+      conv.convert(yname_);
+      conv.convert(yunit_);
+      conv.convert(xname_);
+      conv.convert(xunit_);
+      conv.convert(group_name_);
+      conv.convert(group_comment_);
     }
 
     // get info string

python/IMCtermite.pyx

@@ -5,6 +5,16 @@ from IMCtermite cimport cppimctermite
 
 import json as jn
 import decimal
+import platform
+
+# auxiliary function for codepage conversion
+def get_codepage(chn) :
+    if platform == 'Windows' :
+        chndec = jn.loads(chn.decode(errors="ignore"))
+        chncdp = chndec["codepage"]
+        return 'utf-8' if chncdp is None else chncdp
+    else :
+        return 'utf-8'
 
 cdef class imctermite:
 
@@ -22,7 +32,7 @@ cdef class imctermite:
   # get JSON list of channels
   def get_channels(self, bool include_data):
     chnlst = self.cppimc.get_channels(True,include_data)
-    chnlstjn = [jn.loads(chn.decode(errors="ignore")) for chn in chnlst]
+    chnlstjn = [jn.loads(chn.decode(get_codepage(chn),errors="ignore")) for chn in chnlst]
     return chnlstjn
 
   # print single channel/all channels

python/VERSION

@@ -1 +1 @@
-2.0.9
+2.0.16