This is the mail archive of the libffi-discuss@sourceware.org mailing list for the libffi project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: libffi-3.0.12 : test cls_struct_va1.c false failure?


On Sun, Feb 24, 2013 at 8:16 PM, Dennis Clarke <dclarke@blastwave.org> wrote:
>
>> On Sun, Feb 24, 2013 at 5:42 PM, Dennis Clarke <dclarke@blastwave.org>
>> wrote:
>> >
>> > Dear AG et al :
>> >
>> > OKay, making good progress here on a Niagara T2 server running
>> Solaris 10.
>> >
>> > Firstly we have what looks to me as a false failure.  In the
>> testsuite log I see
>> > this :
>> > .
>> > .
>> > .
>> > spawn /opt/solarisstudio12.3/bin/cc ../../testsuite/libffi.call/cls_struct_va1.c
>> >  -errfmt=error -erroff=%none -errshort=full -xstrconst -xildoff -m64
>> -xmemalign=
>> > 8s -xnolibmil -Xa -xcode=pic32 -xregs=no%appl -xlibmieee -mc -g -xs
>> -ftrap=%none
>> >  -Qy -xbuiltin=%none -xdebugformat=dwarf -xunroll=1 -xtarget=ultraT2
>> -xcache=8/1
>> > 6/4:4096/64/16 -I/usr/local/build/libffi-3.0.12_SunOS5.10_sparcv9.002/sparc-sun-
>> > solaris2.10/testsuite/../include -I../../testsuite/../include -I/usr/local/build
>> > /libffi-3.0.12_SunOS5.10_sparcv9.002/sparc-sun-solaris2.10/testsuite/../include/
>> > .. -L/usr/local/build/libffi-3.0.12_SunOS5.10_sparcv9.002/sparc-sun-solaris2.10/
>> > testsuite/../.libs -lffi -lm -o ./cls_struct_va1.exe^M
>> > PASS: libffi.call/cls_struct_va1.c (test for excess errors)
>> > Setting LD_LIBRARY_PATH to :/usr/local/build/libffi-3.0.12_SunOS5.10_sparcv9.002
>> > /sparc-sun-solaris2.10/testsuite/../.libs::/usr/local/build/libffi-3.0.12_SunOS5
>> > .10_sparcv9.002/sparc-sun-solaris2.10/testsuite/../.libs
>> > spawn [open ...]^M
>> > 4 5 6 10 11 12 13 14 20 21
>> > res: 1054032
>> > PASS: libffi.call/cls_struct_va1.c execution test
>> > FAIL: libffi.call/cls_struct_va1.c output pattern test, is 4 5 6 10
>> 11 12 13 14 20 21
>> > res: 1054032
>> > , should match 4 5 6 10 11 12 13 14 20 21^M?
>> > res: 42
>>
>> Look at the last line for both:
>> res: 1054032
>> res: 42
>>
>> They are not the same are they?  So this is not a false failure.
>
> Let me try a different approach.
>
> Where is that number "1054032" coming from ?

That is coming out from the program via a printf statement.  The value
passed to the printf is wrong.

>
> If there are failures, and these are valid .. why ?
>
> If the result should be "the meaning of life" then why am I getting 1054032 ?

I don't know without anyone else debugging the program.
The closure that is produced gets produces the wrong place for the return value.
>
> The failure isn't telling me much other than "lib did not build right .. good luck".
> Not very useful .

How is that not useful if you are writing in to a development of the
library list?  It might be the case sparc is not as well tested and
nobody has given some time to debug this failure.

The testcase is about closure and struct passing/integer returning.

Thanks,
Andrew Pinski


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]