Mercurial > hg > sidinfo

changeset 278:d73ccb155878

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Implement support for outputting CP850 and CP437 in our simple fallback character set convertor.
author Matti Hamalainen <ccr@tnsp.org>
date Mon, 06 Jan 2020 20:28:55 +0200
parents a23dae1a4849
children d5ab136cdc97
files sidinfo.c
diffstat 1 files changed, 123 insertions(+), 39 deletions(-) [+]
line wrap: on
line diff
--- a/sidinfo.c	Mon Jan 06 18:16:25 2020 +0200
+++ b/sidinfo.c	Mon Jan 06 20:28:55 2020 +0200
@@ -19,11 +19,6 @@
 // Some constants
 //
 
-// Default character encoding to convert to
-// NOTE! Do not change unless you are using iconv()!!
-// The fallback converter does not handle other encodings.
-#define SET_DEF_CHARSET      "utf8"
-
 // HVSC documents directory
 #define SET_HVSC_DOCUMENTS   "DOCUMENTS"
 
@@ -49,6 +44,15 @@
 };
 
 
+enum
+{
+    TH_LANG_UTF8,
+    TH_LANG_ISO88591,
+    TH_LANG_CP850,
+    TH_LANG_CP437,
+};
+
+
 typedef struct
 {
     int cmd;
@@ -130,9 +134,11 @@
 SIDLibSTILDB *sidSTILDB = NULL;
 
 
-BOOL    setUseChConv;
+BOOL    setUseOutConv;
 #ifdef HAVE_ICONV
-iconv_t setChConv;
+iconv_t setIConvCtx;
+#else
+int     setOutLang;
 #endif
 
 
@@ -382,6 +388,25 @@
 }
 
 
+static const uint8_t si_lang_iso88591_to_cp850[16*6] = {
+0xff, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee,
+0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8,
+0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8,
+0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1,
+0x85, 0xa0, 0x83, 0xc6, 0x84, 0x86, 0x91, 0x87, 0x8a, 0x82, 0x88, 0x89, 0x8d, 0xa1, 0x8c, 0x8b,
+0xd0, 0xa4, 0x95, 0xa2, 0x93, 0xe4, 0x94, 0xf6, 0x9b, 0x97, 0xa3, 0x96, 0x81, 0xec, 0xe7, 0x98,
+};
+
+static const uint8_t si_lang_iso88591_to_cp437[16*6] = {
+0xff, 0xad, 0x9b, 0x9c, 0x00, 0x9d, 0x00, 0x00, 0x00, 0x00, 0xa6, 0xae, 0xaa, 0x00, 0x00, 0x00,
+0xf8, 0xf1, 0xfd, 0x00, 0x00, 0xe6, 0x00, 0xfa, 0x00, 0x00, 0xa7, 0xaf, 0xac, 0xab, 0x00, 0xa8,
+0x00, 0x00, 0x00, 0x00, 0x8e, 0x8f, 0x92, 0x80, 0x00, 0x90, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0xa5, 0x00, 0x00, 0x00, 0x00, 0x99, 0x00, 0x00, 0x00, 0x00, 0x00, 0x9a, 0x00, 0x00, 0xe1,
+0x85, 0xa0, 0x83, 0x00, 0x84, 0x86, 0x91, 0x87, 0x8a, 0x82, 0x88, 0x89, 0x8d, 0xa1, 0x8c, 0x8b,
+0x00, 0xa4, 0x95, 0xa2, 0x93, 0x00, 0x94, 0xf6, 0x00, 0x97, 0xa3, 0x96, 0x81, 0x00, 0x00, 0x98,
+};
+
+
 char *siConvertCharset(const char *src)
 {
 #ifdef HAVE_ICONV
@@ -395,40 +420,81 @@
 
     while (srcLeft > 0)
     {
-        size_t ret = iconv(setChConv, &srcPtr, &srcLeft, &outPtr, &outLeft);
+        size_t ret = iconv(setIConvCtx, &srcPtr, &srcLeft, &outPtr, &outLeft);
         if (ret == (size_t) -1)
             break;
     }
 
 #else
-    // Fallback ISO-8859-1 to UTF-8 conversion
-    size_t srcSize = strlen(src),
-           outSize = srcSize * 2 + 1;
+    // Fallback conversion of ISO-8859-1 to X
+    size_t srcSize = strlen(src), outSize, minLeft;
     const uint8_t *srcPtr = (const uint8_t *) src;
+    const uint8_t *tab;
     uint8_t *outBuf, *outPtr;
-    if ((outBuf = outPtr = th_malloc(outSize + 1)) == NULL)
+
+    switch (setOutLang)
+    {
+        case TH_LANG_UTF8:
+            outSize = srcSize * 2;
+            minLeft = 2;
+            break;
+
+        default:
+            outSize = srcSize;
+            minLeft = 1;
+    }
+
+    if ((outBuf = outPtr = th_malloc(outSize)) == NULL)
         return NULL;
 
-    while (srcSize > 0 && outSize >= 2)
+    while (srcSize > 0 && outSize >= minLeft)
     {
-        if (*srcPtr < 0x80)
-        {
-            *outPtr++ = *srcPtr;
-            outSize--;
-        }
-        else
-        if (*srcPtr < 0xBF)
+        switch (setOutLang)
         {
-            *outPtr++ = 0xC2;
-            *outPtr++ = *srcPtr;
-            outSize -= 2;
+            case TH_LANG_UTF8:
+                // Not 100% correct really, but close enough
+                if (*srcPtr < 0x80)
+                {
+                    *outPtr++ = *srcPtr;
+                    outSize--;
+                }
+                else
+                if (*srcPtr < 0xBF)
+                {
+                    *outPtr++ = 0xC2;
+                    *outPtr++ = *srcPtr;
+                    outSize -= 2;
+                }
+                else
+                {
+                    *outPtr++ = 0xC3;
+                    *outPtr++ = *srcPtr - 0x40;
+                    outSize -= 2;
+                }
+                break;
+
+            case TH_LANG_ISO88591:
+                *outPtr++ = *srcPtr;
+                outSize--;
+                break;
+
+            case TH_LANG_CP850:
+            case TH_LANG_CP437:
+                // Not 100% correct either, but close enough
+                tab = (setOutLang == TH_LANG_CP850) ? si_lang_iso88591_to_cp850 : si_lang_iso88591_to_cp437;
+
+                if (*srcPtr < 0x7f)
+                    *outPtr++ = *srcPtr;
+                else
+                if (*srcPtr >= 0xA0)
+                    *outPtr++ = tab[*srcPtr - 0xA0];
+                else
+                    *outPtr++ = '?';
+
+                outSize--;
+                break;
         }
-        else
-        {
-            *outPtr++ = 0xC3;
-            *outPtr++ = (*srcPtr - 0x40);
-            outSize -= 2;
-        }
+
         srcPtr++;
         srcSize--;
     }
@@ -943,11 +1009,11 @@
 static void siPrintPSIDInfoLine(FILE *outFile, BOOL *shown,
     const char *fmt, const int otype,
     const char *d_str, const int d_int,
-    const BOOL useConv)
+    const BOOL convert)
 {
     char *str, *tmp;
 
-    if (setUseChConv && d_str != NULL && useConv)
+    if (setUseOutConv && d_str != NULL && convert)
     {
         char *tmp2 = siConvertCharset(d_str);
         tmp = siEscapeString(tmp2, optEscapeChars);
@@ -1318,15 +1384,33 @@
                 setLang[i++] = th_tolower(*ptr);
         }
         setLang[i] = 0;
-    }
 
 #ifdef HAVE_ICONV
-    // Initialize iconv, check if we have language/charset
-    setChConv = iconv_open(setLang != NULL ? setLang : SET_DEF_CHARSET, "iso88591");
-    setUseChConv = setChConv != (iconv_t) -1;
+        // Initialize iconv, check if we have language/charset
+        setIConvCtx = iconv_open("utf8", "iso88591");
+        setUseOutConv = setIConvCtx != (iconv_t) -1;
 #else
-    setUseChConv = setLang != NULL && strcmp(setLang, SET_DEF_CHARSET) == 0;
+        // Check if we can use our fallback converter
+        if (strcmp(setLang, "utf8") == 0)
+            setOutLang = TH_LANG_UTF8;
+        else
+        if (strcmp(setLang, "iso88591") == 0 ||
+            strcmp(setLang, "cp819") == 0 ||
+            strcmp(setLang, "latin1") == 0 ||
+            strcmp(setLang, "cp28591") == 0)
+            setOutLang = TH_LANG_ISO88591;
+        else
+        if (strcmp(setLang, "cp850") == 0)
+            setOutLang = TH_LANG_CP850;
+        else
+        if (strcmp(setLang, "cp437") == 0)
+            setOutLang = TH_LANG_CP437;
+        else
+            setOutLang = TH_LANG_ISO88591;
+
+        setUseOutConv = setOutLang != TH_LANG_ISO88591;
 #endif
+    }
 
     // Parse command line arguments
     if (!th_args_process(argc, argv, optList, optListN,
@@ -1334,7 +1418,7 @@
         goto out;
 
     THMSG(2, "Requested output LANG='%s', use charset conversion=%s\n",
-        setLang, setUseChConv ? "yes" : "no");
+        setLang, setUseOutConv ? "yes" : "no");
 
     if (optOneLineFieldSep != NULL)
     {
@@ -1483,8 +1567,8 @@
 out:
 
 #ifdef HAVE_ICONV
-    if (setUseChConv)
-        iconv_close(setChConv);
+    if (setUseOutConv)
+        iconv_close(setIConvCtx);
 #endif
 
     th_free(setLang);