Re: ruby build now failing in aarch64

[Michael Tremer <michael.tremer@ipfire.org>]

Hello,

I have no idea why we are running into this problem. Others have too, but there seems to be other circumstances.

I pushed a patch the other day that fixes the problem for us:

  https://git.ipfire.org/?p=ipfire-2.x.git;a=commitdiff;h=e6791a9e4a3210201188daa981d3b2d2c092846e

This is a workaround and hopefully we can soon drop this.

-Michael

> On 15 May 2025, at 17:18, Adolf Belka <adolf.belka@ipfire.org> wrote:
> 
> Hi All,
> 
> On 15/05/2025 16:30, Adolf Belka wrote:
>> Hi All,
>> The fix to udev to not fail if the linker emits warnings solved the udev build on aarch64 for me.
>> However now it has got to building ruby and that has now failed.
>> The error message is
>> In file included from vm_core.h:118,
>>                  from eval_intern.h:5,
>>                  from debug.c:16:
>> vm_callinfo.h: In function 'vm_ci_dump':
>> internal.h:89:72: error: 'RUBY_FUNCTION_NAME_STRING' undeclared (first use in this function)
>>    89 | #define rp(obj) rb_obj_info_dump_loc((VALUE)(obj), __FILE__, __LINE__, RUBY_FUNCTION_NAME_STRING)
>>       |                                                                        ^~~~~~~~~~~~~~~~~~~~~~~~~
>> vm_callinfo.h:183:9: note: in expansion of macro 'rp'
>>   183 |         rp(ci);
>>       |         ^~
>> internal.h:89:72: note: each undeclared identifier is reported only once for each function it appears in
>>    89 | #define rp(obj) rb_obj_info_dump_loc((VALUE)(obj), __FILE__, __LINE__, RUBY_FUNCTION_NAME_STRING)
>>       |                                                                        ^~~~~~~~~~~~~~~~~~~~~~~~~
>> vm_callinfo.h:183:9: note: in expansion of macro 'rp'
>>   183 |         rp(ci);
>>       |         ^~
>> vm_callinfo.h: In function 'vm_ci_new_':
>> internal.h:89:72: error: 'RUBY_FUNCTION_NAME_STRING' undeclared (first use in this function)
>>    89 | #define rp(obj) rb_obj_info_dump_loc((VALUE)(obj), __FILE__, __LINE__, RUBY_FUNCTION_NAME_STRING)
>>       |                                                                        ^~~~~~~~~~~~~~~~~~~~~~~~~
>> vm_callinfo.h:221:16: note: in expansion of macro 'rp'
>>   221 |     if (debug) rp(ci);
>>       |                ^~
>> and the RUBY_FUNCTION_NAME_STRING undeclared I found a comment that it could be related to gcc15 but it seemed to be for windows systems.
>> The version of ruby we have is 3.4.1 but 3.4.4 has been just issued with a fix
>> Bug #21286: Windows - MSYS2 just updated to GCC 15.1.0, builds failing - Ruby - Ruby Issue Tracking System
>> The bug report does have comments that it is strongly impacting windows systems but it is also affecting other OS's.
>> I will try doing a build with the latest ruby and see if that solves the aarch64 build.
> 
> Unfortunately building ruby-3.4.4 did not fix the above problem. The same errors still occurred.
> 
> Regards,
> Adolf.
> 
>> Regards,
>> Adolf.