Re: [Yap-users] yap-6.3 bug report

[Vitor S. C. <vs...@gm...>]

Dear Gong Su

Thanks!

The first bug was called only once, maybe that's how come I didn't notice.
The second bug was not in qlyw.yap, it was a bug in a data-base update that
broke an invariant for qlyw. The bug didn't actually affect execution,
that's why it survived for so long. The updates are at:

https://github.com/vscosta/yap-6.3.git

Don't be worried about the number of commits, they are mostly moving
comments around :)

Great report!

Vitor

On Fri, Sep 12, 2014 at 6:04 PM, Gong Su <go...@ho...> wrote:

> Hi, I'm not sure whether this is the right email for reporting yap bugs.
>
> I'm trying to compile and run, on 64-bit Intel RedHat Linux, the latest
> version of yap-6.3 which I got from the git repository using
>
> git clone git://git.code.sf.net/p/yap/yap-6.3
>
> I found a couple of bugs:
>
> 1. In file C/qlyw.c, function save_int, the local variable v is declared
> as UInt (8 bytes) instead of int (4 bytes). As a result, the save_bytes
> call will only save the first 4 bytes of v. On a little endian machine,
> this happens to be correct. But on a big endian machine, you will get all
> 0s.
>
> 2. When configure yap with the option --enable-threads, you cannot do
> save_program. save_program will result in a segmentation fault like the
> following:
>
> %
> %
> % YAP OOOPS: tried to access illegal address 0x18!!!!.
> %
> %
> %
> % PC: prolog:qsave_program/1 at clause 1
> %   Continuation: prolog:qsave_program/1 at clause 1
> Active Choice-Points:
> prolog:$user_call(save_program(test),save_program(test))
> prolog:$yes_no(save_program(test),?-)
> prolog:$catch(_131266,user:$Error(_131266),_1048524)
> prolog:[Segmentation fault
>
> I have tried to do some debugging and it appears that there are predicate
> structures with garbage FunctorOfPred pointers. I added an fprintf line in
> save_program function:
>
> fprintf(stderr, "%p %x %p\n", pp->FunctorOfPred,
> pp->FunctorOfPred->KindOfPE, pp->FunctorOfPred->NameOfFE);
>
> and here are a few lines (among many):
>
> 0x6a87c0 bb00 0x6a8790
> 0x6a8640 bb00 0x6a8610
> 0x6a84c0 6a84f0 (nil)
> 0x6a8370 6a83a0 0x726f5f
> 0x6a8220 bb00 0x6a8200
> 0x6a80b0 bb00 0x6a8090
> 0x6a7f40 bb00 0x6a7f20
> 0x6a7db0 bb00 0x6a7d80
> 0x6a7c10 bb00 0x6a7bf0
> 0x6a7aa0 6a7ad0 0x70666564
> 0x6a7950 6a7980 0x706665646e
> 0x6a7800 6a7830 0x6d6574737973
> 0x6a76b0 6a76e0 0x6d657473
> 0x6a7540 bb00 0x6a7520
> 0x6a73d0 bb00 0x68b7d0
> 0x695fa0 74c540 0x21
> 0x6a7140 bb00 0x6a7110
> 0x6a6fc0 bb00 0x6a6f90
> 0x6a6e40 bb00 0x6a6e10
> 0x6a6cc0 bb00 0x6a6c90
> 0x6a6b40 bb00 0x6a6b10
> 0x6a69c0 bb00 0x6a6990
>
>
>
>
> ------------------------------------------------------------------------------
> Want excitement?
> Manually upgrade your production database.
> When you want reliability, choose Perforce
> Perforce version control. Predictably reliable.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
> _______________________________________________
> Yap-users mailing list
> Yap...@li...
> https://lists.sourceforge.net/lists/listinfo/yap-users
>
>