This bug exists in the current j9.6 beta, also.
But I'm encountering problems debugging where my debugger is telling
me that the code is doing something which I can observe conflicts with
the behavior of the code.
Specifically, /:~&.#:8+0 3 2 0 1 2 2 0 gives a valid result, but gdb
is skipping lines when I'm single stepping through it:
464 if(n>5){ // TUNE
(gdb)
465 if(t&(C4T+INT+FL)){
(gdb)
475 }
(gdb)
482 if(!z){ // not a special case
(gdb)
486 }
(gdb)
488 if(!z){A t;
(gdb)
498 EPILOG(z);
(gdb)
And I can't get much done with broken tools... I could probably work
around this issue (using printf style debugging), but someone else
with more familiarity with the code would probably make better
progress than I've been doing.
FYI,
--
Raul
On Thu, Dec 28, 2023 at 7:02 PM Raul Miller <[email protected]> wrote:
>
> It's worth noting that this problem exists in the current J release also:
>
> ]A=: #:0 3 2 0 1 2 2 0
> 0 0
> 1 1
> 1 0
> 0 0
> 0 1
> 1 0
> 1 0
> 0 0
> /:~A
> 0 0
> 0 0
> 0 0
> 1 0
> 1 0
> 1 0
> 0 0
> 0 0
> JVERSION
> Engine: j9.5.1/j64avx2/windows
> Build: commercial/2023-12-20T04:09:52/clang-16-0-6/SLEEF=1
> Library: 9.5.5
> Qt IDE: 2.5.2/6.5.3(6.5.3)
> OS Ver: Windows 11 Version 22H2 10.0.22621
> Platform: Win 64
> Installer: j9.5 install
> InstallPath: c:/other/j9.5
> Contact: www.jsoftware.com
>
> Testing suggests that the problem does not occur in j903.
>
> FYI,
>
> --
> Raul
>
> On Thu, Dec 28, 2023 at 4:35 PM Raul Miller <[email protected]> wrote:
> >
> > Odd, this is now working for me on 9.4.3. I do not know why it was
> > failing earlier , but it works now. (Previously, I was getting what
> > looked like a repeat of
> > https://www.jsoftware.com/pipermail/source/2023-October/001539.html
> > and my attempted workaround also failed.) Hopefully, this will
> > continue to work. I don't know what I changed between the earlier
> > failed attempt and this successful build.
> >
> > Under j902, I get:
> >
> > cc -c -o am.o ../../../../jsrc/am.c -Werror -fPIC -O2
> > -fvisibility=hidden -fwrapv -fno-strict-aliasing -Wextra
> > -Wno-unused-parameter -Wno-sign-compare -Wno-clobbered -Wno-empty-body
> > -Wno-unused-value -Wno-pointer-sign -Wno-parentheses -Wno-type-limits
> > -Wno-maybe-uninitialized -Wno-shift-negative-value
> > -Wno-implicit-fallthrough -Wno-cast-function-type -DC_AVX=1 -mavx
> > ../../../../jsrc/am.c: In function ‘jtcasev’:
> > ../../../../jsrc/am.c:114:1: error: function may return address of
> > local variable [-Werror=return-local-addr]
> > 114 | } /* z=:b}x0,x1,x2,...,x(m-2),:x(m-1) */
> > | ^
> > In file included from ../../../../jsrc/j.h:970,
> > from ../../../../jsrc/am.c:6:
> > ../../../../jsrc/am.c:80:15: note: declared here
> > 80 | fauxblockINT(aafaux,4,1);
> > | ^~~~~~
> > ../../../../jsrc/ja.h:360:31: note: in definition of macro ‘fauxblockINT’
> > 360 | #define fauxblockINT(z,n,r) I z[(AKXR(r)>>LGSZI)+(n)] //
> > define a block, big enough to hold n atoms at rank r, for use in
> > fauxINT
> > | ^
> > cc1: all warnings being treated as errors
> > make: *** [makefile-libj:166: am.o] Error 1
> >
> > Perhaps I should avoid j902, but older versions can be handy when
> > researching changes.
> >
> > Thanks,
> >
> > --
> > Raul
> >
> >
> > On Thu, Dec 28, 2023 at 9:36 AM bill lam <[email protected]> wrote:
> > >
> > > On linux or Mac, it should be as simple as typing make2/build_all.sh
> > > What are the errors that you got?
> > >
> > > On Thu, Dec 28, 2023 at 10:23 PM Raul Miller <[email protected]>
> > > wrote:
> > >
> > > > I decided I should try taking a look at this... but I'm running into
> > > > numerous errors trying to compile j9.4.3 (or, for that matter, j902).
> > > >
> > > > So... I guess I won't try to look into this. But I figure I should
> > > > give people a heads up that at least one machine has issues with
> > > > compiling J.
> > > >
> > > > --
> > > > Raul
> > > >
> > > > On Thu, Dec 28, 2023 at 5:40 AM bill lam <[email protected]> wrote:
> > > > >
> > > > > I got a similar result. I believe this bug is related to the boolean
> > > > > argument.
> > > > > coercing the array to integer can work correctly.
> > > > >
> > > > > p =. 00 + 8 2 $ 0 0 1 1 1 0 0 0 0 1 1 0 1 0 0 0
> > > > >
> > > > > On Thu, Dec 28, 2023 at 6:02 PM Marcin Żołek <
> > > > > [email protected]> wrote:
> > > > >
> > > > > > Hello,
> > > > > > I was trying to run my code on j9.4 Linux, j9.4 MacOS, j9.5 MacOS
> > > > > > and
> > > > it
> > > > > > crashes.
> > > > > >
> > > > > > In terminal
> > > > > > /Applications/j9.5/bin/jconsole bug.ijs < input.txt
> > > > > > JE has crashed, likely due to an internal bug. Please report the
> > > > > > code
> > > > > > which caused the crash, as well as the following printout, to the J
> > > > > > programming forum.
> > > > > > Could not generate stack trace: no debug info in Mach-O executable
> > > > > > (-1)
> > > > > > ...
> > > > > >
> > > > > >
> > > > -----------------------------------------------------------------------------
> > > > > > 5 3 1 1[1] 6692 abort /Applications/j9.5/bin/jconsole
> > > > > > bug.ijs <
> > > > > > input.txt
> > > > > >
> > > > > > bug.ijs:
> > > > > > ":@((#/.~ +/@, ~.)@/:~)@(".;._2)&.stdin''
> > > > > >
> > > > > > input.txt:
> > > > > > 0 0
> > > > > > 1 1
> > > > > > 1 0
> > > > > > 0 0
> > > > > > 0 1
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 0 0
> > > > > > I figured out that the problem is with sorting:
> > > > > > p =. 8 2 $ 0 0 1 1 1 0 0 0 0 1 1 0 1 0 0 0
> > > > > > p
> > > > > > 0 0
> > > > > > 1 1
> > > > > > 1 0
> > > > > > 0 0
> > > > > > 0 1
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 0 0
> > > > > > /:~ p NB. returns various and wrong results
> > > > > > 0 0
> > > > > > 0 0
> > > > > > 0 0
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 1 1
> > > > > > 1 1
> > > > > >
> > > > > > or
> > > > > >
> > > > > > 0 0
> > > > > > 0 0
> > > > > > 0 0
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 1 0
> > > > > > 0 0
> > > > > > 0 0
> > > > > > ~.@:/:~ p NB. crashes jqt
> > > > > >
> > > > > > Best regards,
> > > > > > Marcin
> > > > > > ----------------------------------------------------------------------
> > > > > > For information about J forums see
> > > > > > http://www.jsoftware.com/forums.htm
> > > > > >
> > > > > ----------------------------------------------------------------------
> > > > > For information about J forums see http://www.jsoftware.com/forums.htm
> > > > ----------------------------------------------------------------------
> > > > For information about J forums see http://www.jsoftware.com/forums.htm
> > > >
> > > ----------------------------------------------------------------------
> > > For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
