Re: [head tinderbox] failure on ia64/ia64

On Fri, Sep 26, 2008 at 12:42 PM, Marcel Moolenaar <xcllnt_at_mac.com> wrote:
>
> On Sep 26, 2008, at 12:19 PM, Peter Wemm wrote:
>
>> On Fri, Sep 26, 2008 at 8:05 AM, Marcel Moolenaar <xcllnt_at_mac.com> wrote:
>>>
>>> On Sep 26, 2008, at 5:56 AM, Kostik Belousov wrote:
>>>
>>>> On Fri, Sep 26, 2008 at 08:40:25AM -0400, FreeBSD Tinderbox wrote:
>>>>>
>>>>> TB --- 2008-09-26 12:27:35 - tinderbox 2.3 running on
>>>>> freebsd-current.sentex.ca
>>>>> TB --- 2008-09-26 12:27:35 - starting HEAD tinderbox run for ia64/ia64
>>>>> TB --- 2008-09-26 12:27:35 - cleaning the object tree
>>>>> TB --- 2008-09-26 12:28:10 - cvsupping the source tree
>>>>> TB --- 2008-09-26 12:28:10 - /usr/bin/csup -r 3 -g -L 1 -h localhost -s
>>>>> /tinderbox/HEAD/ia64/ia64/supfile
>>>>> TB --- 2008-09-26 12:28:17 - building world (CFLAGS=-O -pipe)
>>>>> TB --- 2008-09-26 12:28:17 - cd /src
>>>>> TB --- 2008-09-26 12:28:17 - /usr/bin/make -B buildworld
>>>>>>>>
>>>>>>>> World build started on Fri Sep 26 12:28:19 UTC 2008
>>>>>>>> Rebuilding the temporary build tree
>>>>>>>> stage 1.1: legacy release compatibility shims
>>>>>>>> stage 1.2: bootstrap tools
>>>>>>>> stage 2.1: cleaning up the object tree
>>>>>>>> stage 2.2: rebuilding the object tree
>>>>>>>> stage 2.3: build tools
>>>>>>>> stage 3: cross tools
>>>>>>>> stage 4.1: building includes
>>>>>>>> stage 4.2: building libraries
>>>>>
>>>>> [...]
>>>>> building static gcc library
>>>>> ranlib libgcc.a
>>>>> sh /src/tools/install.sh -C -o root -g wheel -m 444   libgcc_eh.a
>>>>> /obj/ia64/src/tmp/usr/lib
>>>>> sh /src/tools/install.sh -C -o root -g wheel -m 444   libgcc.a
>>>>> /obj/ia64/src/tmp/usr/lib
>>>>> ===> gnu/lib/csu (obj,depend,all,install)
>>>>> cc -O -pipe  -DIN_GCC -DHAVE_LD_EH_FRAME_HDR -DDT_CONFIG -D__GLIBC__=3
>>>>> -finhibit-size-directive -fno-inline-functions  -fno-exceptions
>>>>> -fno-zero-initialized-in-bss  -fno-zero-initialized-in-bss
>>>>> -fno-toplevel-reorder
>>>>> -I/src/gnu/lib/csu/../../../contrib/gcclibs/include
>>>>> -I/src/gnu/lib/csu/../../../contrib/gcc/config
>>>>> -I/src/gnu/lib/csu/../../../contrib/gcc -I.
>>>>> -I/src/gnu/lib/csu/../../usr.bin/cc/cc_tools -x assembler-with-cpp
>>>>> -include
>>>>> osreldate.h -std=gnu89  -g0 -DCRT_BEGIN  -c -o crtbegin.o
>>>>> /src/gnu/lib/csu/../../../contrib/gcc/config/ia64/crtbegin.asm
>>>>> /obj/ia64/src/tmp/usr/include/osreldate.h: Assembler messages:
>>>>> /obj/ia64/src/tmp/usr/include/osreldate.h:38: Error: Unknown opcode
>>>>> `extern int getosreldate(void)'
>>>>> *** Error code 1
>>>>>
>>>>> Stop in /src/gnu/lib/csu.
>>>>> *** Error code 1
>>>>
>>>> May be the right location for getosreldate() prototype is unistd.h then.
>>>> Do you agree ?
>>>
>>> I just fixed it by adding -D_OSRELDATE_H to CFLAGS. This prevents the
>>> prototype from being declared while still defining __FreeBSD_version.
>>> In theory -- I haven't tested it yet...
>>
>> The question is..  how is __STDC__ getting set for asm code?  It most
>> definitely is not "standard C" code.
>
> gcc (read: cc1) defines it:
>
> hob% cc -E -x assembler-with-cpp -dM /dev/null | grep STDC
> #define __STDC_HOSTED__ 1
> #define __STDC__ 1
>
> If you added the __STDC__ conditional especially to avoid
> build breakage on ia64, then we should probably remove it.

I originally added the __STDC__ thing to stop imake breaking.  It was
a long time ago though when I last messed with it.

-- 
Peter Wemm - peter_at_wemm.org; peter_at_FreeBSD.org; peter_at_yahoo-inc.com; KI6FJV
"All of this is for nothing if we don't go to the stars" - JMS/B5
"If Java had true garbage collection, most programs would delete
themselves upon execution." -- Robert Sewell