search

MichaelChirico / r-bugs

A ⚠️read-only⚠️mirror of https://bugs.r-project.org/
20 stars 0 forks source link

[BUGZILLA #1630] build fails dumping R code in 'methods' on i686-pc-linux-gnu #1378

Closed MichaelChirico closed 4 years ago

MichaelChirico commented 4 years ago

From: "George N. White III" <aa056@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> Using the same settings in config.site as were used for R-1.4.1, R-1.5.0 fails to build.

./configure --enable-R-shlib

Using defaults and using PPro optimizations, the build failed in package methods:

dumping R code in package 'methods' gmake[4]: *** [../../../library/methods/R/all.rda] Error 139 gmake[4]: Leaving directory `/mnt/Data/src/install/R-1.5.0/src/library/methods'

The command that fails is:

../../../bin/R --vanilla --verbose < all.R

The same failure occurs using defaults for CFLAGS, etc.

Install patches:

zcat R-release.diff.gz | patch -p1

R is now configured for i686-pc-linux-gnu

Source directory: . Installation directory: /usr/local C compiler: gcc -D__NO_MATH_INLINES -mieee-fp -O2 \ -pipe -s -mpentiumpro -mcpu=pentiumpro -march=pentiumpro -mieee-fp \ -malign-loops=2 -malign-jumps=2 -malign-functions=2 \ -mpreferred-stack-boundary=2 C++ compiler: g++ -mieee-fp -O2 -pipe -s -mpentiumpro \ -mcpu=pentiumpro -march=pentiumpro -mieee-fp -malign-loops=2 \ -malign-jumps=2 -malign-functions=2 -mpreferred-stack-boundary=2 FORTRAN compiler: g77 -O2 -pipe -s -mpentiumpro \ -mcpu=pentiumpro -march=pentiumpro -mieee-fp -malign-loops=2 \ -malign-jumps=2 -malign-functions=2 -mpreferred-stack-boundary=2 X11 support: yes Gnome support: no Tcl/Tk support: yes Readline support: yes

R profiling support: yes R as a shared library: yes

The same error occurs using the patched version.

$ gdb ../../../bin/R.bin core GNU gdb 4.18 Copyright 1998 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i586-mandrake-linux"... Core was generated by `/mnt/Data/src/install/R-1.5.0/bin/R.bin --vanilla --slave'. Program terminated with signal 11, Segmentation fault. Cannot access memory at address 0x40015238

0 0x403cb10d in ?? ()Cannot access memory at address 0x40015238

(gdb)

### The following code is only executed when dumping
assign(".saveImage", FALSE, .GlobalEnv)
.First.lib("methods", "methods", .GlobalEnv)

$ ../../../bin/R --debugger=gdb

(gdb) run --vanilla < all.R [...]

Program received signal SIGSEGV, Segmentation fault. 0x403cb10d in Rf_install () from /mnt/Data/src/install/R-1.5.0/bin/libR.so (gdb) where

0 0x403cb10d in Rf_install () from /mnt/Data/src/install/R-1.5.0/bin/libR.so

1 0x404f5a85 in R_initMethodDispatch ()

from /mnt/Data/src/install/R-1.5.0/library/methods/libs/methods.so

2 0x807c6a2 in do_dotCode () at eval.c:88

3 0x808b032 in Rf_eval () at eval.c:88

4 0x808beae in do_begin () at eval.c:88

5 0x808aeaf in Rf_eval () at eval.c:88

6 0x808b3f0 in Rf_applyClosure () at eval.c:88

7 0x808b0aa in Rf_eval () at eval.c:88

8 0x80a0d8d in R_ReplConsole () at eval.c:88

9 0x80a140a in run_Rmainloop () at eval.c:88

10 0x80a1421 in Rf_mainloop () at eval.c:88

11 0x80f5f6d in main () at eval.c:88

12 0x400f38a6 in __libc_start_main (main=0x80f5f58
, argc=2,

ubp_av=0xbffff354, init=0x804a564 <_init>, fini=0x81677a4 <_fini>,
rtld_fini=0x4000b980 <_dl_fini>, stack_end=0xbffff34c)
at ../sysdeps/generic/libc-start.c:129

(gdb)

The following information is from R-1.4.1 big.report()

--please do not edit the information below--

Version: platform = i686-pc-linux-gnu arch = i686 os = linux-gnu system = i686, linux-gnu status = major = 1 minor = 4.1 year = 2002 month = 01 day = 30 language = R

Search Path: .GlobalEnv, package:ctest, Autoloads, package:base

-- George N. White III <aa056@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> Halifax, NS, Canada

METADATA

MichaelChirico commented 4 years ago

From: George White <aa056@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> On Tue, 4 Jun 2002, John Chambers wrote:

aa056@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> wrote:
> 
> Using the same settings in config.site as were used for R-1.4.1,
> R-1.5.0 fails to build.
> 
> ./configure --enable-R-shlib
> 
> Using defaults and using PPro optimizations, the build failed in
> package methods:
> 
> dumping R code in package 'methods'
> gmake[4]: *** [../../../library/methods/R/all.rda] Error 139
> gmake[4]: Leaving directory
`/mnt/Data/src/install/R-1.5.0/src/library/methods'
> 
> The command that fails is:
> 
>     ../../../bin/R --vanilla --verbose < all.R
> 
> The same failure occurs using defaults for CFLAGS, etc.
> 
> Install patches:
> 
>   zcat R-release.diff.gz | patch -p1
> 
> R is now configured for i686-pc-linux-gnu
> 
>   Source directory:          .
>   Installation directory:    /usr/local
>   C compiler:                gcc  -D__NO_MATH_INLINES -mieee-fp -O2 \
>  -pipe -s -mpentiumpro -mcpu=pentiumpro -march=pentiumpro -mieee-fp \
>  -malign-loops=2 -malign-jumps=2 -malign-functions=2 \
>  -mpreferred-stack-boundary=2
>   C++ compiler:              g++  -mieee-fp -O2 -pipe -s  -mpentiumpro \
>  -mcpu=pentiumpro -march=pentiumpro -mieee-fp -malign-loops=2 \
>  -malign-jumps=2 -malign-functions=2 -mpreferred-stack-boundary=2
>   FORTRAN compiler:          g77  -O2 -pipe -s -mpentiumpro \
> -mcpu=pentiumpro -march=pentiumpro -mieee-fp -malign-loops=2 \
> -malign-jumps=2 -malign-functions=2 -mpreferred-stack-boundary=2
>   X11 support:               yes
>   Gnome support:             no
>   Tcl/Tk support:            yes
>   Readline support:          yes
> 
>   R profiling support:       yes
>   R as a shared library:     yes
> 
> The same error occurs using the patched version.
> 

I couldn't reproduce this with the --enable-R-shlib in the
configuration, on either of my current Linux versions (Redhat 7.1, YDL
2.2); e.g.,

I had no problem building the default configuration. Comparing the resulting Makeconf with the one for the failed build I found that configure did not set MAIN_LDFLAGS to --export-dynamic when using my config.site (which has the MAIN_LDFLAGS entry commented out).

If I edit Makeconf and rebuild R.bin the build finished and "make check" passes.

-- George White <aa056@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> Halifax, Nova Scotia

METADATA

MichaelChirico commented 4 years ago

NOTES: report on 1.5.0

METADATA

MichaelChirico commented 4 years ago

Audit (from Jitterbug): Wed Jun 5 09:49:39 2002 ripley moved from incoming to not-reproducible Thu May 18 08:47:17 2006 ripley changed notes

METADATA