Svadev mailing list

[Jingyue Wu <wujingyue AT gmail.com>]

Thanks for fixing this issue! I don't think "RUNNO" was picked up on my side. But when "dsaopt" (which is the last command) returns 0, "XFAIL" still treats it as an error. I think this is why the test case failed. But anyway, since you removed this test case already, we needn't worry about it.

dsa/regression/2012-04-29.SQLiteBUInfiniteRecursion.ll is still failing on my side. All other test cases are passed though. The test case keeps eating memory and never stops. gdb seems to suggest DSA is running into infinite recursions (a very long call stack). I'll send you the gdb call stack when I get a chance. Shall I issue a separate bug report?

My platform: 

Linux 3.2

gcc 4.6

llvm 3.0

Jingyue

On Fri, Jun 1, 2012 at 9:31 AM, Will Dietz <willdtz AT gmail.com> wrote:

Very sorry for how long this is taking to resolve.

Not sure why lit is picking up "RUNNO:" for you, as far as I know
"RUNNO" was just something I made up: equivalent to commenting out a
line of code, I sometimes change RUN: lit comands to "RUNNO".

Regardless, that command is never meant to be actually run (the diff),
so I'll just remove it.

Are you seeing any other issues?  Just wondering, what platform are
you building on?

~Will

On Wed, May 30, 2012 at 12:11 PM, Jingyue Wu <wujingyue AT gmail.com> wrote:
> I did "make" from the test directory. I tried "make lit" just now and the
> problem still occurs.
>
> I looked at print.c, and am wondering whether "XFAIL: *" conflicts with
> "RUNNO: diff". "XFAIL" expects the result to be non-zero, but "RUNNO" will
> return zero if the two outputs are different. Am I right?
>
> Jingyue
>
> On Tue, May 29, 2012 at 11:34 PM, Will Dietz <willdtz AT gmail.com> wrote:
>>
>> This test is "supposed" to print different pointer values, but the way
>> the lit test is constructed this should not be causing you failures.
>>
>> How are you running the tests?  The supported/expected way is "make
>> lit" from the test directory of your object tree.
>>
>> ~Will
>>
>> On Thu, May 10, 2012 at 9:24 AM, Jingyue Wu <wujingyue AT gmail.com> wrote:
>> > print.c.tmp.refout:
>> >
>> > @F 1234 1234 123.22 123.45
>> > @TEST ID3:AAAUnknown
>> > @TEST ID3:AAAUnknown 0x209f910
>> >
>> > print.c.tmp.refout:
>> >
>> > @F 1234 1234 123.22 123.45
>> > @TEST ID3:AAAUnknown
>> > @TEST ID3:AAAUnknown 0x159b8d0
>> >
>> > Jingyue
>> >
>> >
>> > On Wed, May 9, 2012 at 10:03 PM, Jingyue Wu <wujingyue AT gmail.com> wrote:
>> >>
>> >> The errors I see:
>> >>
>> >> /home/jingyue/Research/llvm/llvm-3.0/utils/lit/lit.py
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test --path
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/Release+Asserts/bin -s -v
>> >> XPASS: PoolAlloc/DSA :: dsa/var_arg/print.c (120 of 133)
>> >> ******************** TEST 'PoolAlloc/DSA :: dsa/var_arg/print.c' FAILED
>> >> ********************
>> >> Script:
>> >> --
>> >> /home/jingyue/Research/llvm/llvm-3.0/Release+Asserts/bin/clang -O0
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/print.c -S
>> >> -emit-llvm -o - |
>> >> /home/jingyue/Research/llvm/llvm-3.0/Release+Asserts/bin/llvm-as >
>> >>
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/Output/print.c.tmp.bc
>> >> /home/jingyue/Research/llvm/llvm-3.0/Release+Asserts/bin/lli
>> >>
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/Output/print.c.tmp.bc
>> >> >
>> >>
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/Output/print.c.tmp.refout
>> >> dsaopt
>> >>
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/Output/print.c.tmp.bc
>> >> -ds-aa -O3 -o - |
>> >> /home/jingyue/Research/llvm/llvm-3.0/Release+Asserts/bin/lli >
>> >>
>> >> /home/jingyue/Research/llvm/poolalloc-trunk/test/dsa/var_arg/Output/print.c.tmp.out
>> >> --
>> >> Exit Code: 0
>> >>
>> >> ********************
>> >>                          -- Testing: 133 tests, 4 threads --
>> >>
>> >>
>> >>
>> >> On Wed, May 9, 2012 at 9:36 PM, Will Dietz <willdtz AT gmail.com> wrote:
>> >>>
>> >>> Oh, those test/dsa/var_arg/*.c tests are terrible (at least in their
>> >>> present format).
>> >>>
>> >>> Strange.  Well, certainly don't worry about them.  I'm not sure what
>> >>> the right thing to do with them is.  They're marked XFAIL, so strange
>> >>> that you're seeing them give errors... :/.
>> >>>
>> >>> Could you forward me the errors you're seeing, if you get a chance?
>> >>>
>> >>> Thanks!
>> >>>
>> >>> ~Will
>> >>>
>> >>> On Wed, May 9, 2012 at 8:29 PM, Jingyue Wu <wujingyue AT gmail.com>
>> >>> wrote:
>> >>> > svn up'ed, and there are still errors on dsa/var_arg/print.c.
>> >>> >
>> >>> > Jingyue
>> >>> >
>> >>> >
>> >>> > On Wed, May 9, 2012 at 9:14 PM, Will Dietz <willdtz AT gmail.com>
>> >>> > wrote:
>> >>> >>
>> >>> >> Yes, sorry.  I've always been hesitant to mark known failures as
>> >>> >> XFAIL
>> >>> >> (I suppose I'm worried they'll get lost that way), but I've gone
>> >>> >> ahead
>> >>> >> and done so to make it clear what's known/expected to fail and
>> >>> >> what's
>> >>> >> not :).
>> >>> >>
>> >>> >> We have bugzilla, after all.
>> >>> >>
>> >>> >> Sorry about that!
>> >>> >>
>> >>> >> ~Will
>> >>> >>
>> >>> >> On Wed, May 9, 2012 at 8:06 PM, Jingyue Wu <wujingyue AT gmail.com>
>> >>> >> wrote:
>> >>> >> > I checked out the latest trunk version, and found it failed with
>> >>> >> > dsa/callgraph/loop1.ll and dsa/callgraph/loop2.ll. Does this
>> >>> >> > happen
>> >>> >> > to
>> >>> >> > you?
>> >>> >> >
>> >>> >> > Jingyue
>> >>> >> >
>> >>> >> > On Wed, May 9, 2012 at 7:46 PM, John Criswell
>> >>> >> > <criswell AT illinois.edu>
>> >>> >> > wrote:
>> >>> >> >>
>> >>> >> >> On 5/9/12 5:03 PM, Will Dietz wrote:
>> >>> >> >>>
>> >>> >> >>> Hi Jingyue, taking this off bugzilla since I believe this has
>> >>> >> >>> gone
>> >>> >> >>> sufficiently off-topic, certainly for that particular bug
>> >>> >> >>> report
>> >>> >> >>> :).
>> >>> >> >>>
>> >>> >> >>> On Wed, May 9, 2012 at 4:05 PM,<bugzilla-daemon AT llvm.org>
>> >>> >> >>>  wrote:
>> >>> >> >>> [snip]
>> >>> >> >>>
>> >>> >> >>>
>> >>> >> >>> FWIW I believe John was suggesting you write some lit tests for
>> >>> >> >>> DS-AA,
>> >>> >> >>> not trying to run the test-suite (although that can be useful
>> >>> >> >>> too).
>> >>> >> >>
>> >>> >> >>
>> >>> >> >> Yes, I was suggesting that there be lit tests for ds-aa to help
>> >>> >> >> ensure
>> >>> >> >> that it keeps working.
>> >>> >> >>
>> >>> >> >> -- John T.
>> >>> >> >>
>> >>> >> >
>> >>> >> >
>> >>> >> >
>> >>> >> > --
>> >>> >> > Jingyue Wu
>> >>> >> > Department of Computer Science
>> >>> >> > Columbia University
>> >>> >> > New York, NY 10027
>> >>> >> > http://www.cs.columbia.edu/~jingyue
>> >>> >
>> >>> >
>> >>> >
>> >>> >
>> >>> > --
>> >>> > Jingyue Wu
>> >>> > Department of Computer Science
>> >>> > Columbia University
>> >>> > New York, NY 10027
>> >>> > http://www.cs.columbia.edu/~jingyue
>> >>
>> >>
>> >>
>> >>
>> >> --
>> >> Jingyue Wu
>> >> Department of Computer Science
>> >> Columbia University
>> >> New York, NY 10027
>> >> http://www.cs.columbia.edu/~jingyue
>> >
>> >
>> >
>> >
>> > --
>> > Jingyue Wu
>> > Department of Computer Science
>> > Columbia University
>> > New York, NY 10027
>> > http://www.cs.columbia.edu/~jingyue
>
>
>
>
> --
> Jingyue Wu
> Department of Computer Science
> Columbia University
> New York, NY 10027
> http://www.cs.columbia.edu/~jingyue

--
Jingyue Wu
Department of Computer Science
Columbia University
New York, NY 10027
http://www.cs.columbia.edu/~jingyue