I've just built R-devel (r60889) from source on my Mac OS X 10.6.8 using gcc from Xcode and gfortran from Simon. I didn't use all the recommended configuration flags only ./configure && make Still, I seem to pass "make check" and "make test-Segfault" in the tests directory with no problem, but the following leads to a persistent error + instant abort. 1) Start R 2) Type "tools::" 3) Tab to attempt to auto-complete resulting in some variation on mweylandt-laptop:~ michaelweylandt$ rDEVEL/bin/R --vanilla -q> tools::R(44436) malloc: *** error for object 0x1006d8700: incorrect checksum for freed object - object was probably modified after being freed.*** set a breakpoint in malloc_error_break to debug Abort trap at which point hitting <RET> drops me back into bash. This seems to only occur with tools and no other packages (that I've found). It also occurs whether tools has been library()'d or not, but sometimes it takes two tabbings. My compiler info follows mweylandt-laptop:~ michaelweylandt$ gcc --version i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3) Copyright (C) 2007 Free Software Foundation, Inc. mweylandt-laptop:~ michaelweylandt$ gfortran --version GNU Fortran (GCC) 4.2.1 (Apple Inc. build 5664) Copyright (C) 2007 Free Software Foundation, Inc. I do think that it's only a problem with my local build, as I can't replicate with Simon's nightly builds (http://r.research.att.com/), which are at r60887 currently. Is there anything else I should check with my setup? I don't see any advice on this particular issue in the R installation manual, but I easily could have overlooked it. Thanks, Michael
On 07-10-2012, at 17:34, R. Michael Weylandt wrote:> I've just built R-devel (r60889) from source on my Mac OS X 10.6.8 > using gcc from Xcode and gfortran from Simon. I didn't use all the > recommended configuration flags only > > ..... > My compiler info follows > > mweylandt-laptop:~ michaelweylandt$ gcc --version > i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3) > Copyright (C) 2007 Free Software Foundation, Inc. > > mweylandt-laptop:~ michaelweylandt$ gfortran --version > GNU Fortran (GCC) 4.2.1 (Apple Inc. build 5664) > Copyright (C) 2007 Free Software Foundation, Inc. > > I do think that it's only a problem with my local build, as I can't > replicate with Simon's nightly builds (http://r.research.att.com/), > which are at r60887 currently. Is there anything else I should check > with my setup? I don't see any advice on this particular issue in the > R installation manual, but I easily could have overlooked it.Just a thought. According to http://r.research.att.com/tools/ shouldn't you be using gcc 5664 on Snow Leopard. If I read the tools page correctly you seem to be using a Lion version of gcc. On my 10.6.8 I have [.../~]: gcc --version i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5664) Copyright (C) 2007 Free Software Foundation, Inc. [.../~]: gfortran --version GNU Fortran (GCC) 4.2.1 (Apple Inc. build 5664) Copyright (C) 2007 Free Software Foundation, Inc. Berend
I would check the version of libreadline you compiled against. Tab completion involves that library. It works correctly for all my local builds, and nothing in that area in the R sources has changed for a long time. I believe I am using http://r.research.att.com/libs/readline-5.2-12-darwin8-bin4.tar.gz . And this is on Lion whereas Simon is building on Leopard (AFAIK). On 07/10/2012 16:34, R. Michael Weylandt wrote:> I've just built R-devel (r60889) from source on my Mac OS X 10.6.8 > using gcc from Xcode and gfortran from Simon. I didn't use all the > recommended configuration flags only > > ./configure && make > > Still, I seem to pass "make check" and > "make test-Segfault" in the tests directory with no problem, but the > following leads > to a persistent error + instant abort. > > 1) Start R > 2) Type "tools::" > 3) Tab to attempt to auto-complete > > resulting in some variation on > > mweylandt-laptop:~ michaelweylandt$ rDEVEL/bin/R --vanilla -q >> tools::R(44436) malloc: *** error for object 0x1006d8700: incorrect checksum for freed object - object was probably modified after being freed. > *** set a breakpoint in malloc_error_break to debug > Abort trap > > at which point hitting <RET> drops me back into bash. This seems to > only occur with tools and no other packages (that I've found). It also > occurs whether tools has been library()'d or not, but sometimes it > takes two tabbings. > > My compiler info follows > > mweylandt-laptop:~ michaelweylandt$ gcc --version > i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3) > Copyright (C) 2007 Free Software Foundation, Inc. > > mweylandt-laptop:~ michaelweylandt$ gfortran --version > GNU Fortran (GCC) 4.2.1 (Apple Inc. build 5664) > Copyright (C) 2007 Free Software Foundation, Inc. > > I do think that it's only a problem with my local build, as I can't > replicate with Simon's nightly builds (http://r.research.att.com/), > which are at r60887 currently. Is there anything else I should check > with my setup? I don't see any advice on this particular issue in the > R installation manual, but I easily could have overlooked it. > > Thanks, > Michael > > ______________________________________________ > R-devel at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel >-- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595