Re: "A diagram of C23 basic types"

On 4/12/2025 10:21 AM, David Brown wrote:

On 12/04/2025 07:43, Lawrence D'Oliveiro wrote:

On Thu, 10 Apr 2025 11:37:30 +0300, Michael S wrote:
>

So, IMHO, if C waits for C++ then it will wait forever.

>
Seems like C is already committed to avoiding incompatibilities with C++,
if the decision on thousands separators in numbers is anything to go by.

 It has been committed to avoiding /unnecessary/ incompatibilities with C++ since C++ perhaps C99.  The C and C++ committees will not introduce similar but incompatible features to each other unless there is very good reason to do so.  The fact that some people think 1_000 is nicer than 1'000 is not nearly enough to pick a different type of digit separator for C.
 

My main annoyance is that ' is already used for character literals, so say:
0x123'A'BCD
The ' needs to be handled as special in the tokenizing, otherwise it would give 3 tokens 0x123 'A' BCD. Its use is also inconsistent with its use as a character literal.
Then again, maybe one could argue the same for '_', but at least:
* Already used by a number of other languages;
* Serves as a common spacer in identifiers;
* ...
Personally though, BGBCC supports both.
* Potentially, _ could be disallowed in C++ mode if needed, as the tokenizer is language aware (partly due to subtle differences in token rules).