RFR: 8365296: Build failure with Clang due to -Wformat warning after JDK-8364611 [v3]

[Albert Mingkun Yang]

On Thu, 14 Aug 2025 16:02:54 GMT, Guanqiang Han <ghan at openjdk.org> wrote:

>> This issue is related to the definition of __sigset_t (__sigset_t  act.sa_mask).   In the glibc source, __sigset_t is defined in multiple places:  
>> - bits/types/__sigset_t.h 
>> <img width="1071" height="415" alt="image02" src="https://github.com/user-attachments/assets/4dff7546-f0b3-448c-832f-54f7b9ffc476" />
>> 
>> 
>> - sysdeps/unix/sysv/linux/bits/types/__sigset_t.h (introduced after glibc 2.25)   
>> <img width="1059" height="487" alt="image01" src="https://github.com/user-attachments/assets/54101f9b-3b1a-4943-83ee-6f6471c060a6" />
>> 
>> 
>> During compilation with Clang, the latter definition appears to be used, where __sigset_t is a struct, causing the printf("%X", act.sa_mask) to fail.
>> 
>> we can detect whether _SIGSET_NWORDS is defined and handle the printing differently based on that, ensuring correct handling of the struct type.
>
> Guanqiang Han has updated the pull request incrementally with two additional commits since the last revision:
> 
>  - Update exePrintSignalDisposition.c
>    
>    a small fix
>  - Update exePrintSignalDisposition.c
>    
>    Refactor act.sa_mask output

The fix looks reasonable. However, the discrepancy btw and AIX and others seems unnecessary complexity. I'd prefer having `ifdef` removed also.

Looking at how these prints are used/tested, I believe `sa_mask` is not really needed and it doesn't seem very useful also. Therefore, I'd suggest removing it, just like what's done for AIX. My 2c.

-------------

PR Review: https://git.openjdk.org/jdk/pull/26771#pullrequestreview-3122060281