We need to define source and execution chracter sets (see section 5.2.1 of the C11 standard) for SDCC.
The easiest approach would be to support only the basic character set as required by the standard. Currently, we have some (but not complete) support for iso-8859-1. Here's a suggestion for supporting a bit more: The execution character set and encoding are the same as the source character set and encoding. The source file shall be written in an character set and encoding, for which the following properties hold: 1) Every character from the basic character set is encoded as a value in the range 0 to 127, which is the same as in the execution character set of the compiler used to compile SDCC. 2) Every multibyte character that contains a byte of value in the range 0 to 127 consists of exactly one byte. However, unless the source is written in UTF-8-encoed unicode, the following restrictions to standard-compliance apply: 3) The value of universal character constants is undefined. 4) The value of u8-prefixed string literals is undefined. 5) The use of u- and U-prefixed string literals results in undefined behaviour. This should be easy to implement, will allow to make use of most common character sets one might want to use on an embedded device today. And for today's de-facto standard, UTF-8-encoded Unicode, we get full standard-compliance. Philipp
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------------
_______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
