Input and output code pages

• It’s possible to get the Win32 compiler to be in a weird state where it starts writing UTF-8 encoded bytes for ASCII string data. This seems to depend on the order of #pragma code_page directives in a file and/or the code page given on the CLI.
  • A Windows-1252 encoded file that has only 1 RCDATA { "Ó" } in it will (with all default settings) be compiled into the single byte 0xD3 (Ó in Windows-1252). If the option /c65001 is given to the CLI, it will instead be compiled into 0xEF 0xBF 0xBD, the UTF-8 sequence for U+FFFD (the replacement character).
  • A Windows-1252 encoded file that has 1 RCDATA { "Ó" } with #pragma code_page(1252) before it will be compiled into the single byte 0xD3 (Ó in Windows-1252). If the option /c65001 is given to the CLI, it will instead be compiled into 0xC3 0x93, the UTF-8 sequence for Ó.
    • The /c65001 behavior can be ‘counteracted’ if there is any #pragma code_page before the #pragma code_page(1252) (it will compile into 0xD3 again).
  • A Windows-1252 encoded file that has 1 RCDATA { "Ó" } with #pragma code_page(65001) before it will be compiled into the single byte 0x3F (the ? character). If the option /c65001 is given to the CLI, it will instead be compiled into 0xEF 0xBF 0xBD, the UTF-8 sequence for U+FFFD (the replacement character).
    • The /c65001 behavior can be triggered without using the /c option if there is any #pragma code_page before the #pragma code_page(65001)
  • It seems like there are actually two distinct settings: input code page and output code page. The first #pragma code_page does not affect the output code page, but all the rest do. The /c CLI option affects both.