[NSIS-commits] SF.net SVN: nsis:[7431] NSIS/trunk

[<and...@us...>]

Revision: 7431
          http://sourceforge.net/p/nsis/code/7431
Author:   anders_k
Date:     2024-05-31 18:34:18 +0000 (Fri, 31 May 2024)
Log Message:
-----------
Support setting source file decoding charset with Python-style PEP 263 magic comment

Modified Paths:
--------------
    NSIS/trunk/Docs/src/history.but
    NSIS/trunk/Docs/src/script.but
    NSIS/trunk/Source/lineparse.cpp
    NSIS/trunk/Source/lineparse.h
    NSIS/trunk/Source/script.cpp
    NSIS/trunk/Source/utf.cpp

Modified: NSIS/trunk/Docs/src/history.but
===================================================================
--- NSIS/trunk/Docs/src/history.but	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Docs/src/history.but	2024-05-31 18:34:18 UTC (rev 7431)
@@ -12,6 +12,8 @@
 
 \S2{} Minor Changes
 
+\b Support setting source file decoding charset with magic comment (\W{http://sf.net/p/nsis/patches/292}{patch #292})
+
 \b Make System/Resource.dll reproducible (\W{http://sf.net/p/nsis/patches/312}{patch #312})
 
 \S2{} Translations

Modified: NSIS/trunk/Docs/src/script.but
===================================================================
--- NSIS/trunk/Docs/src/script.but	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Docs/src/script.but	2024-05-31 18:34:18 UTC (rev 7431)
@@ -31,6 +31,10 @@
 
 If you want a parameter to start with ; or # put it in quotes.
 
+If the first or second line in the file is a \#{pep263}comment with the following format, the rest of the file is decoded using the specified character set.
+
+\c # -*- coding: utf-8 -*-
+
 \\<b\\>Plug-ins\\</b\\>
 
 To call a plug-in, use 'plugin::command [parameters]'. For more info see \R{plugindlls}{Plug-in DLLs}.

Modified: NSIS/trunk/Source/lineparse.cpp
===================================================================
--- NSIS/trunk/Source/lineparse.cpp	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Source/lineparse.cpp	2024-05-31 18:34:18 UTC (rev 7431)
@@ -52,11 +52,11 @@
   return m_incommentblock;
 }
 
-int LineParser::parse(TCHAR *line, int ignore_escaping/*=0*/) // returns -1 on error
+int LineParser::parse(const TCHAR *line, int ignore_escaping/*=0*/, NStreamEncoding*pEncChange) // returns -1 on error
 {
   freetokens();
   bool bPrevCB=m_incommentblock;
-  int n=doline(line, ignore_escaping);
+  int n=doline(line, ignore_escaping, pEncChange);
   if (n) return n;
   if (m_nt)
   {
@@ -305,8 +305,14 @@
   m_nt=0;
 }
 
-int LineParser::doline(TCHAR *line, int ignore_escaping/*=0*/)
+static bool IsPEP263EncodingCharacter(TCHAR c)
 {
+  TCHAR low = S7ChLwr(c);
+  return ('a' <= low && low <= 'z') || ('0' <= c && c <= '9') || c == '.' || c == '-' || c == '_';
+}
+
+int LineParser::doline(const TCHAR *line, int ignore_escaping/*=0*/, NStreamEncoding*pEncChange)
+{
   m_nt=0;
   m_incomment = false;
   while (*line == _T(' ') || *line == _T('\t')) line++;
@@ -332,6 +338,21 @@
       if (*line == _T(';') || *line == _T('#'))
       {
         m_incomment = true;
+        TCHAR *p = const_cast<TCHAR*>(line);
+        if (pEncChange && *++p) // Check for Python PEP 263 magic encoding comment
+        {
+            p = _tcsstr(p, L"coding");
+            if (p) p += sizeof("coding") - 1; else break;
+            if (*p == ':' || *p == '=') ++p; else break;
+            while (*p == _T(' ') || *p == _T('\t')) ++p;
+            TCHAR buf[200];
+            if (strtrycpy(buf, p, COUNTOF(buf)))
+            {
+                for (p = buf; IsPEP263EncodingCharacter(*p);) ++p;
+                *p = '\0';
+                pEncChange->SetCodepage(GetEncodingFromString(buf));
+            }
+        }
         break;
       }
       if (*line == _T('/') && *(line+1) == _T('*'))
@@ -346,7 +367,7 @@
         else if (*line == _T('`')) lstate=4;
         if (lstate) line++;
         int nc=0;
-        TCHAR *p = line;
+        const TCHAR *p = line;
         while (*line)
         {
           if (line[0] == _T('$') && line[1] == _T('\\'))

Modified: NSIS/trunk/Source/lineparse.h
===================================================================
--- NSIS/trunk/Source/lineparse.h	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Source/lineparse.h	2024-05-31 18:34:18 UTC (rev 7431)
@@ -20,6 +20,7 @@
 #define _LINEPARSE_H_
 
 #include "tchar.h"
+#include "utf.h"
 
 class LineParser {
   public:
@@ -29,7 +30,7 @@
 
     bool inComment();
     bool inCommentBlock();
-    int parse(TCHAR *line, int ignore_escaping=0); // returns -1 on error
+    int parse(const TCHAR *line, int ignore_escaping=0, NStreamEncoding*pEncChange=NULL); // returns -1 on error
     int getnumtokens();
     void eattoken();
     int gettoken_int(int token, int *success=0) const;
@@ -48,7 +49,7 @@
   private:
 
     void freetokens();
-    int doline(TCHAR *line, int ignore_escaping=0);
+    int doline(const TCHAR *line, int ignore_escaping=0, NStreamEncoding*pEncChange=NULL);
     inline int validate_token_index(int token, int *success=0) const;
 
     int m_eat;

Modified: NSIS/trunk/Source/script.cpp
===================================================================
--- NSIS/trunk/Source/script.cpp	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Source/script.cpp	2024-05-31 18:34:18 UTC (rev 7431)
@@ -219,7 +219,7 @@
     LineParser prevline(inside_comment);
     prevline.parse((TCHAR*)m_linebuild.get());
     LineParser thisline(inside_comment);
-    thisline.parse((TCHAR*)str);
+    thisline.parse(str);
 
     if (prevline.inComment() && !thisline.inComment())
     {
@@ -242,8 +242,14 @@
   // escaped quotes should be ignored for compile time commands that set defines
   // because defines can be inserted in commands at a later stage
   bool ignore_escaping = (!_tcsnicmp((TCHAR*)m_linebuild.get(),_T("!define"),7) || !_tcsncicmp((TCHAR*)m_linebuild.get(),_T("!insertmacro"),12));
-  res=line.parse((TCHAR*)m_linebuild.get(), ignore_escaping);
 
+  NStreamEncoding enc(NStreamEncoding::UNKNOWN);
+
+  res=line.parse((TCHAR*)m_linebuild.get(), ignore_escaping, linecnt < 3 ? &enc : NULL);
+
+  if (enc.GetCodepage() != NStreamEncoding::UNKNOWN && curlinereader)
+      curlinereader->StreamEncoding().SafeSetCodepage(enc.GetCodepage());
+
   inside_comment = line.inCommentBlock();
 
   // if ignoring, ignore all lines that don't begin with an exclamation mark

Modified: NSIS/trunk/Source/utf.cpp
===================================================================
--- NSIS/trunk/Source/utf.cpp	2024-05-30 19:07:00 UTC (rev 7430)
+++ NSIS/trunk/Source/utf.cpp	2024-05-31 18:34:18 UTC (rev 7431)
@@ -274,6 +274,7 @@
   }
   return 0;
 }
+
 UINT DetectUTFBOM(FILE*strm)
 {
   /*\
@@ -325,19 +326,32 @@
   return 0;
 }
 
+static bool IsUTFEncodingString(const TCHAR*s, const TCHAR*suff)
+{
+  if (S7ChLwr(s[0]) == 'u' && S7ChLwr(s[1]) == 't' && S7ChLwr(s[2]) == 'f')
+  {
+    s += 3 + (s[3] == '-' || s[3] == '_');
+    return !_tcsicmp(s, suff);
+  }
+  return false;
+}
+
 WORD GetEncodingFromString(const TCHAR*s, bool&BOM)
 {
   BOM = false;
-  if (!_tcsicmp(s,_T("ACP"))) return NStreamEncoding::ACP;
-  if (!_tcsicmp(s,_T("OEM"))) return NStreamEncoding::OEMCP;
-  if (!_tcsicmp(s,_T("UTF8"))) return NStreamEncoding::UTF8;
-  if ((!_tcsicmp(s,_T("UTF8SIG")) || !_tcsicmp(s,_T("UTF8BOM"))) && (BOM = true))
+  if (!_tcsicmp(s,_T("ACP")))
+    return NStreamEncoding::ACP;
+  if (!_tcsicmp(s,_T("OEM")))
+    return NStreamEncoding::OEMCP;
+  if (IsUTFEncodingString(s, _T("8")))
     return NStreamEncoding::UTF8;
-  if (!_tcsicmp(s,_T("UTF16LE")) || (!_tcsicmp(s,_T("UTF16LEBOM")) && (BOM = true)))
+  if (IsUTFEncodingString(s, _T("8SIG")) || IsUTFEncodingString(s, _T("8BOM")))
+    return (BOM = true, NStreamEncoding::UTF8);
+  if (IsUTFEncodingString(s, _T("16LE")) || (IsUTFEncodingString(s, _T("16LEBOM")) && (BOM = true)))
     return NStreamEncoding::UTF16LE;
-  if (!_tcsicmp(s,_T("UTF16BE")) || (!_tcsicmp(s,_T("UTF16BEBOM")) && (BOM = true)))
-    return NStreamEncoding::UTF16BE;
-  if (S7IsChEqualI('C',*s++) && S7IsChEqualI('P',*s++))
+  if (IsUTFEncodingString(s, _T("16BE")) || (IsUTFEncodingString(s, _T("16BEBOM")) && (BOM = true)))
+    return NStreamEncoding::UTF16LE;
+  if (S7IsChEqualI('C', *s++) && S7IsChEqualI('P', *s++))
   {
     int cp = _tstoi(s);
     if (cp > 0 && cp < NStreamEncoding::CPCOUNT) return (WORD) cp;
@@ -344,6 +358,7 @@
   }
   return NStreamEncoding::UNKNOWN;
 }
+
 WORD GetEncodingFromString(const TCHAR*s)
 {
   bool bom;
@@ -365,10 +380,10 @@
   case UTF8: p = _T("UTF8"); break;
   case BINARY: p = _T("BIN"); break;
   default: 
-    _stprintf(mybuf,_T("CP%u"),CP);
+    _stprintf(mybuf,_T("CP%u"), CP);
     if (CP >= NStreamEncoding::CPCOUNT) p = _T("?");
   }
-  _tcscpy(Buf,p);
+  _tcscpy(Buf, p);
 }
 
 bool NBaseStream::Attach(FILE*hFile, WORD enc, bool Seek /*= true*/)

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.