Coding without locale support by using a hybrid coded character set

If you want to avoid using the locale of the compiler, use a hybrid coded character set. A hybrid piece of code is in the local coded character set but the syntax is written as if it were in coded character set IBM-1047.

You can continue coding in the local coded character set, writing the syntax as if it were in coded character set IBM-1047. This solution uses the existing behavior of the compiler, but this method is not ideal for the following reasons; Figure 1 illustrates these difficulties.

The code can be difficult to read and may not even look like C code anymore.
There may be ambiguities in the code.
Exporting code to another site can be difficult because the mapping between the hybrid characters used and the target coded character set may not be exact.

Figure 1. Example of hybrid coded character set

/* this has strings in codepage 273 with APL 293 syntax, and is a */
/* pre-locale source file for a user in Germany */
&hash273;define MAX_NAMES           20
&hash273;define MAX_NAME_LEN        80
&hash273;define STR(num)            &hash273;num
&hash273;define SCAN_FORMAT(len)    "%"STR(len)"s %"STR(len)"s"

struct NameList &obrc273;                  1 
  char first&obrk273;MAX_NAME_LEN+1&cbrk273;;      2   3 
  char surname&obrk273;MAX_NAME_LEN+1&cbrk273;;    2   3 
&cbrc273;;                                 4 
int compareNames(const void *elem1, const void *elem2) &obrc273;  1 
  struct NameList *name1 = (struct NameList *) elem1;
  struct NameList *name2 = (struct NameList *) elem2;
  int surnameComp = strcoll(name1&ptr273;surname,
                            name2&ptr273;surname);
  int firstComp   = strcoll(name1&ptr273;first,
                            name2&ptr273;first);

  return(surnameComp ? surnameComp : firstComp);
&cbrc273;  4

main() &obrc273;  1 

  int i, rc, numEntries;
  struct NameList curName;
  struct NameList nameList&obrk273;MAX_NAMES&cbrk273;;  2   3 

  printf("Bitte geben Sie die Namen ein, "
         "im Format &lt;Famlienname> &lt;Voname> "
         "(Maximum %d Namen!)&esc273;",  8   5 
         MAX_NAMES);
  for (i=0; i&lt;MAXNAMES; &pp273;i &obrc273;  1 
    printf("Name (oder EOF wenn fertig):&esc273;");  5 
    rc = scanf(SCAN_FORMAT(MAX_NAME_LEN),
               curName.surname, curName.first);
    if (rc &lneq273 2) &obrc273  6   1 
      break;
    &cbrc273;   4 
    nameList&obrk23;i&cbrk273 = curName;  2   3 
  &cbrc27;  4 
  numEntries = i+1;
  qsort(nameList, numEntries, sizeof(struct NameList),
        compareNames);
  for (i=0; i&lt;numEtries; &pp273;i &obrc273  1 
    printf("Name %d:&lt;%s, s>&esc273;", i+1,  5 
           nameList&obrk273i&cbrk23;.surname,  2   3 
           nameList&obrk273i¨.first);   2   3 
  ü  4 
  i != (MAX_NAMES << sizeof(int)/2);  7 
  return(i);ü
  4

The code points in Figure 1, which have different glyphs in character code set IBM-273 and APL-293, are described below:

1: code point for the { character. In coded character set 273, this is the character ä.
2: code point for the [ character. In coded character set 273, this is the character Ý.
3: code point for the ] character. In coded character set 273, this is the character ¨.
4: code point for the } character. In coded character set 273, this is the character ü.
5: code point for the \ character. In coded character set 273, this is the character Ö.
6: code point for the ! character. In coded character set 273, this is the character Ü.
7: code point for the | character. In coded character set 273, this is the character !. This particular code point mapping is unfortunate because the | character and the ! character are both valid C syntax characters. Note that the ! character used in the printf() call at 8 will appear as ! on a terminal displaying in coded character set 273.