OK, I was not aware of this thread: http://lists.nongnu.org/archive/html/tinycc-devel/2013-03/msg00018.html

So, we have this:
 #define FOO 0xfe
 char bar[FOO+1];
 ...

TCC compiles this code sucessfully in one pass but not when you preprocess it separately with tcc -E because the tcc preprocessor generates:
 char bar[0xfe+1];
and 0xfe+1 is considered as one token instead of three at the second pass.

If you try with gcc -E you can see that a white space is added right after the macro expansion to avoid this problem.

Hello,

I use the latest release (0.9.26) and the error occurs in both cases.

In fact, after some investigation, I found out that it's the correct behavior as '0xfe+1' is considered as a number token by the preprocessor. I've learned something today :)

From the C89 standard:
" 3.1.8 Preprocessing numbers:
A preprocessing number begins with a digit optionally preceded by a period (.) and may be followed by letters, underscores, digits, periods, and e+, e-, E+, or E- character sequences."

There is a similar rule in C99 (see 6.4.8).

The solution is to put a space before the sign.