From 0cdcaea437622a3394ce11266ec9be1ddeb18d0c Mon Sep 17 00:00:00 2001
From: Jean-Francois Dockes <jf@dockes.org>
Date: Thu, 22 Nov 2018 17:41:43 +0100
Subject: [PATCH] mimeparse: use cp1252 instead of iso-8859 on values with
 residual 8bit chars. Also: comments and missing std:: qualifiers

---
 src/common/cstr.h       | 46 +++++++++++++++++++++++++----------------
 src/internfile/Filter.h |  7 +++----
 src/utils/mimeparse.cpp |  2 +-
 3 files changed, 32 insertions(+), 23 deletions(-)

diff --git a/src/common/cstr.h b/src/common/cstr.h
index 17dead32..0b6660d7 100644
--- a/src/common/cstr.h
+++ b/src/common/cstr.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2011 J.F.Dockes
+/* Copyright (C) 2011-2018 J.F.Dockes
  *   This program is free software; you can redistribute it and/or modify
  *   it under the terms of the GNU General Public License as published by
  *   the Free Software Foundation; either version 2 of the License, or
@@ -27,7 +27,6 @@
 // when they are used in a fast loop or are shared.
 
 #include <string>
-using std::string;
 
 // The following slightly hacky preprocessing directives and the
 // companion code in the cpp file looks complicated, but it just
@@ -35,9 +34,9 @@ using std::string;
 // extern declaration and the definition.
 #ifdef RCLIN_CSTR_CPPFILE
 #undef DEF_CSTR
-#define DEF_CSTR(NM, STR) const string cstr_##NM(STR)
+#define DEF_CSTR(NM, STR) const std::string cstr_##NM(STR)
 #else
-#define DEF_CSTR(NM, STR) extern const string cstr_##NM
+#define DEF_CSTR(NM, STR) extern const std::string cstr_##NM
 #endif
 
 DEF_CSTR(caption, "caption");
@@ -63,24 +62,35 @@ DEF_CSTR(fldhtm, "\007");
 DEF_CSTR(wildSpecStChars, "*?[");
 DEF_CSTR(regSpecStChars, "(.[{");
 
-// Values used as keys inside Dijon::Filter::metaData[]. This structure is 
-// used to store all data generated by format-translating filters. It is
-// different from Rcl::Doc for mostly historical reasons. The translation
-// from Filter to Doc occurs inside internfile.cpp
-DEF_CSTR(dj_keyds, "description");
-DEF_CSTR(dj_keyfn, "filename");
-DEF_CSTR(dj_keymd, "modificationdate");
+// Values used as keys inside Dijon::Filter::metaData[].
+
+// The document data.
+DEF_CSTR(dj_keycontent, "content");
+
+// These fields go from the topmost handler (text/plain) into the
+// Rcl::Doc::meta, possibly with some massaging. 
+DEF_CSTR(dj_keyanc, "rclanc");
 DEF_CSTR(dj_keyorigcharset, "origcharset");
+DEF_CSTR(dj_keyds, "description");
+DEF_CSTR(dj_keyabstract, "abstract");
+
+// Built or inherited along the handler stack, then copied to doc
+DEF_CSTR(dj_keyipath, "ipath");
+DEF_CSTR(dj_keyfn, "filename");
+DEF_CSTR(dj_keyauthor, "author");
+DEF_CSTR(dj_keymd, "modificationdate");
+// charset and mimetype are explicitely blocked from going into the doc meta
+DEF_CSTR(dj_keycharset, "charset");
+DEF_CSTR(dj_keymt, "mimetype");
+
+// All other meta fields are directly copied from
+// Dijon::Filter::metaData to Rcl::Doc::meta. The defininitions which
+// follow are just for well-known names, with no particular processing
+// in internfile.
 DEF_CSTR(dj_keytitle, "title");
 DEF_CSTR(dj_keyrecipient, "recipient");
 DEF_CSTR(dj_keymsgid, "msgid");
-DEF_CSTR(dj_keyabstract, "abstract");
-DEF_CSTR(dj_keyauthor, "author");
-DEF_CSTR(dj_keycharset, "charset");
-DEF_CSTR(dj_keycontent, "content");
-DEF_CSTR(dj_keyipath, "ipath");
 DEF_CSTR(dj_keymd5, "md5");
-DEF_CSTR(dj_keymt, "mimetype");
-DEF_CSTR(dj_keyanc, "rclanc");
+
 
 #endif /* _CSTR_H_INCLUDED_ */
diff --git a/src/internfile/Filter.h b/src/internfile/Filter.h
index 35c5e584..c125efa7 100644
--- a/src/internfile/Filter.h
+++ b/src/internfile/Filter.h
@@ -174,10 +174,9 @@ protected:
     /// The MIME type handled by the filter.
     std::string m_mimeType;
 
-    /// Current Metadata dictionary. For multi-document files,
-    /// this may be rebuilt for each sub-document. See
-    /// common/cstr.h for the common key definitions. The document
-    /// text is "content"
+    /// Current Metadata dictionary. For multi-document files, this
+    /// may be rebuilt for each sub-document. See common/cstr.h for
+    /// the common key definitions.
     std::map<std::string, std::string> m_metaData;
 
 };
diff --git a/src/utils/mimeparse.cpp b/src/utils/mimeparse.cpp
index 876b29d6..8417d400 100644
--- a/src/utils/mimeparse.cpp
+++ b/src/utils/mimeparse.cpp
@@ -577,7 +577,7 @@ bool rfc2047_decode(const std::string& in, std::string &out)
     }
 
     if (value.length() > 0) {
-        transcode(value, utf8, "ISO-8859-1", "UTF-8");
+        transcode(value, utf8, "CP1252", "UTF-8");
         out += utf8;
         value.clear();
     }