October 2004 - PPL-devel - BUGSENG Mailman

Bug in ppl_new_LinExpression_with_dimension
by Pedro Vasconcelos 16 Oct '04

16 Oct '04

While experimenting with the Haskell PPL binding, I noticed that the C interface function "ppl_new_LinExpression_with_dimension" always creates an expression with one plus the specified dimension. The following code illustrates the problem: #include <ppl_c.h> main () { ppl_LinExpression_t e; ppl_dimension_type dim_in=3, dim_out;; ppl_initialize(); ppl_new_LinExpression_with_dimension(&e, dim_in); dim_out = ppl_LinExpression_space_dimension(e); printf("%d, %d\n", dim_in, dim_out); /* 3, 4 rather than 3,3 as expected */ ppl_finalize(); } Looking at the source code interface/C/ppl_c.cc, it becomes clear that the expression created has a single Variable(d), therefore dimension d+1: int ppl_new_LinExpression_with_dimension(ppl_LinExpression_t* ple, ppl_dimension_type d) try { *ple = to_nonconst(new LinExpression(0*Variable(d))); /* <----- should be Variable(d-1) */ return 0; } CATCH_ALL This is of course trivial to correct, but I thought it best to report so that it gets corrected in the newer releases. Best regards, Pedro

2 2

ppl NEWS
by Roberto Bagnara 16 Oct '04

16 Oct '04

CVSROOT: /cvs/ppl Module name: ppl Changes by: roberto(a)cs.unipr.it 2004-10-16 23:28:02 Modified files: . : NEWS Log message: Explain why functions in the C interface that compute (space) dimensions no longer return their result. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/NEWS.diff?cvsroot=ppl&r1=1.11…

1 0

ppl NEWS
by Roberto Bagnara 16 Oct '04

16 Oct '04

CVSROOT: /cvs/ppl Module name: ppl Changes by: roberto(a)cs.unipr.it 2004-10-16 23:25:39 Modified files: . : NEWS Log message: Mention the bug fixed in ppl_new_LinExpression_with_dimension(). Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/NEWS.diff?cvsroot=ppl&r1=1.11…

1 0

w3ppl/htdocs/Credits Credits.raw
by Roberto Bagnara 16 Oct '04

16 Oct '04

CVSROOT: /cvs/ppl Module name: w3ppl Changes by: roberto(a)cs.unipr.it 2004-10-16 23:16:29 Modified files: htdocs/Credits : Credits.raw Log message: A couple of consistency improvements. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/w3ppl/htdocs/Credits/Credits.raw.…

1 0

w3ppl/htdocs/Credits Credits.raw
by Roberto Bagnara 16 Oct '04

16 Oct '04

CVSROOT: /cvs/ppl Module name: w3ppl Changes by: roberto(a)cs.unipr.it 2004-10-16 23:12:16 Modified files: htdocs/Credits : Credits.raw Log message: Added Pedro Vasconcelos to the list of users who deserve credit. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/w3ppl/htdocs/Credits/Credits.raw.…

1 0

ppl/interfaces/C ppl_c.cc
by Roberto Bagnara 16 Oct '04

16 Oct '04

CVSROOT: /cvs/ppl Module name: ppl Changes by: roberto(a)cs.unipr.it 2004-10-16 22:56:36 Modified files: interfaces/C : ppl_c.cc Log message: Fixed a bug in ppl_new_LinExpression_with_dimension() reported by Pedro Vasconcelos. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/interfaces/C/ppl_c.cc.diff?cv…

1 0

Re: Haskell PPL bug fix
by Axel Simon 15 Oct '04

15 Oct '04

Roberto, Simon, this is a message is about a Haskell binding to the Parma Polyhedra Library. I had trouble with segfaults which were GC-related. Pedro figured out what I had done wrong: the Haskell compiler GHC replaces the memory allocator in GMP with one that allocates from the Haskell heap. Pedro managed to link the PPL with a private version of GMP but that is not a feasible solution in the long term. The purpose of the message: for Simon M: conceptually it would be nicer if GHC uses it's private version of GMP. Alternatively, GHC could just use malloc for allocating memory and a finalizer each time an Integer goes out of scope. Would that be too slow? for Roberto: this is the reason why there is still no Haskell binding for the PPL. A big thanks to Pedro for figuring this out, Axel. P.S.: More comments below. On Fri, 2004-10-15 at 10:39, Pedro Vasconcelos wrote: > Hello again Axel, > > I've done some more experiments using your PPL binding and I came across > a bug which I think might be related to the situation you've described: > the compiled program will crash misteriously (with a seg fault) after > executing fine for some time. It will happen with repeated solving of > small problems, which led me to believe it is some form of memory > corruption due to GC or deallocation. > > I did a google search and found an related thread in the ghc users > archive. Simon Marlow points out that the shared use of GMP by an > externall library and GHC's runtime system is likely to run into > trouble: > > http://www.mail-archive.com/glasgow-haskell-users@haskell.org/msg06533.html Wow, I don't remember reading that message. Good find. I am amazed why GHC replaces the GMP allocation functions to allocate from the Haskell heap, but since that is what happens, obviously it can only go wrong. Thanks for tracking this down. > None of the solutions suggested in the thread are particularly > straightforward, but Kevin and I came up with an easier fix: what we > require is to link PPL with it's own private GMP library and let the > Haskell RTS use the standard shared GMP. It is sufficient to rename the > symbols defined in GMP by adding a prefix (in linux this can be done > using the "objcopy" util). > > For each symbol XXX defined in libgmp.a, we rename it to myXXX; the > result is libmygmp.a. Apply the same renaming to libppl.a, getting > libmyppl.a using a private GMP library. I have only managed to do this > for the static libraries, but that is good enough for my needs. > > After doing this I have not been able to reproduce the bug, so it looks > like this does indeed solve the problem! That is quite a hack, indeed. I think the better solution is to rename GHC's GMP library. Maybe we should suggest this to Simon. > I'm attaching a small tcl script that builds the renaming table and > creates new versions of the all the required libs. The original libs are > left unchanged. I'll try that when I'll use the PPL from Haskell again. Thanks for tracking this down, Axel.

2 1

RE: Haskell PPL bug fix
by Simon Marlow 15 Oct '04

15 Oct '04

On 15 October 2004 10:25, Axel Simon wrote: > for Simon M: conceptually it would be nicer if GHC uses it's private > version of GMP. Yes. This has been discussed on the GHC User's list (recently I believe). Ideally we'd like to replace GMP with something else. > Alternatively, GHC could just use malloc for > allocating memory and a finalizer each time an Integer goes out of > scope. Would that be too slow? Yes, that would be too slow. Cheers, Simon

1 0

ppl/src C_Polyhedron.defs.hh C_Polyhedron.inli ...
by Enea Zaffanella 15 Oct '04

15 Oct '04

CVSROOT: /cvs/ppl Module name: ppl Changes by: zaffanella(a)cs.unipr.it 2004-10-15 08:48:39 Modified files: src : C_Polyhedron.defs.hh C_Polyhedron.inlines.hh NNC_Polyhedron.defs.hh NNC_Polyhedron.inlines.hh Polyhedron.defs.hh Polyhedron.inlines.hh Polyhedron_chdims.cc Polyhedron_nonpublic.cc max_space_dimension.hh Log message: Removed C_Polyhedron::max_space_dimension() and NNC_Polyehdron::max_space_dimension(): the method Polyhedron::max_space_dimension() no longer uses them and it is not hided by them. Changed all places where we check for dimensions overflows accordingly. Exception throwers made protected to allow calling them from subclasses. The BoundingBox template is now required to have a default constructor. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/C_Polyhedron.defs.hh.diff… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/C_Polyhedron.inlines.hh.d… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/NNC_Polyhedron.defs.hh.di… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/NNC_Polyhedron.inlines.hh… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron.defs.hh.diff?c… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron.inlines.hh.dif… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron_chdims.cc.diff… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/Polyhedron_nonpublic.cc.d… http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/src/max_space_dimension.hh.di…

1 0

ppl/tests addspacedims10.cc
by Enea Zaffanella 13 Oct '04

13 Oct '04

CVSROOT: /cvs/ppl Module name: ppl Changes by: zaffanella(a)cs.unipr.it 2004-10-13 14:19:08 Modified files: tests : addspacedims10.cc Log message: Comment corrected. Patches: http://www.cs.unipr.it/cgi-bin/cvsweb.cgi/ppl/tests/addspacedims10.cc.diff?…

1 0