Parser doesn't recognize _Static_assert() and _Bool

[DanielGibson]

Describe the bug

I imported the Linux kernel source as described in https://github.com/eclipse-cdt/cdt/tree/main/FAQ#whats-the-best-way-to-set-up-the-cdt-to-navigate-linux-kernel-source
(by the way, indexing took like 20min even with an SSD and a fast PC)

I noticed lots of "Syntax errors" that turned out to be caused by Macros that use _Static_assert() or usage of the bool type, which <linux/types.h> defines like typedef _Bool bool; - when looking at that file Eclipse complains about the _Bool.

I could reproduce the _Static_assert() problem in other C and C++ projects, but another C project did not complain about _Bool - not sure why, is there a project-specific setting for the assumed C version that I missed?

To Reproduce
Steps to reproduce the behavior:

Import the Linux source as described in your FAQ and look at drivers/input/serio/i8042.c, maybe expand the used Macros.

Or try _Static_assert(1, "asd"); and _Bool b; this in any other C project, I guess?

Expected behavior

Both _Static_assert(cond, "message") and _Static_assert(cond) should be recognized (and probably ignored).
Same for _Bool, probably regardless of used language version (let the compiler complain about it).

Screenshots

Version Information (please complete the following information):

• OS and OS Version/extra details: XUbuntu Linux 24.04, x86_64
• Eclipse Version from Help -> About: 2025-12 (4.38.0)
• CDT Version from Help -> About -> Installation Details -> Features tab: 12.4.0.202603041756

[DanielGibson]

I just realized: Using _Bool worked in another project, because in the Linux source the typedef _Bool bool; from the header somehow breaks using _Bool and bool in the source.
Maybe the parser thinks that bool is also a compiler-internal type and thus typedefing it is bad?

In that case the parser would be wrong, at least for C - maybe because it's in a header it thinks that it could also be C++ and is thus invalid? I don't know, unfortunately Eclipse only says "Syntax Error" without any further information:

Related: It's not happy with how the kernel defines true and false, either:

[DanielGibson]

If anyone runs into the same problems, my workarounds:

Click to show patch

diff --git a/include/linux/stddef.h b/include/linux/stddef.h
index 80b6bfb944f0..d68d8b650a40 100644
--- a/include/linux/stddef.h
+++ b/include/linux/stddef.h
@@ -7,10 +7,20 @@
 #undef NULL
 #define NULL ((void *)0)
 
+#ifdef __CDT_PARSER__
+// the CDT parser doesn't recognize (and complains about) _Static_assert(),
+// so create a macro of that name with arbitrary arguments
+// (_Static_assert() can have one or two) that does nothing
+#define _Static_assert(...)
+// it also has problems with the enum below, defines for true/false seem to work better
+#define true 1
+#define false 0
+#else
 enum {
        false   = 0,
        true    = 1
 };
+#endif
 
 #undef offsetof
 #define offsetof(TYPE, MEMBER) __builtin_offsetof(TYPE, MEMBER)
diff --git a/include/linux/types.h b/include/linux/types.h
index 7e71d260763c..4d51e809b250 100644
--- a/include/linux/types.h
+++ b/include/linux/types.h
@@ -31,7 +31,12 @@ typedef __kernel_timer_t     timer_t;
 typedef __kernel_clockid_t     clockid_t;
 typedef __kernel_mqd_t         mqd_t;
 
+#ifdef __CDT_PARSER__
+// CDTs parser doesn't like the typedef, but accepts this define
+#define bool _Bool
+#else
 typedef _Bool                  bool;
+#endif
 
 typedef __kernel_uid32_t       uid_t;
 typedef __kernel_gid32_t       gid_t;