R devel - Apr 2003 - Re: Bug#187537: r-base: FTBFS: hangs in tcltk test (PR#2730)

Thanks for the bug report. I will pass it on to R Core.

On Thu, Apr 03, 2003 at 11:01:23PM -0800, Daniel Schepler
wrote:
> Package: r-base
> Version: 1.6.2.cvs.20030403-1
> Severity: serious
> 
> Whenever I try to build r-base using pbuilder, it gets stuck at the
> following point:
> 
> ...
> collecting examples for package 'tcltk' ...
> make[6]: Entering directory
`/tmp/buildd/r-base-1.6.2.cvs.20030403/src/library'
>  >>> Building/Updating help pages for package 'tcltk'
>      Formats: text example
> make[6]: Leaving directory
`/tmp/buildd/r-base-1.6.2.cvs.20030403/src/library'
> running code in 'tcltk-Ex.R' ...
> 
> Part of the output of ps aufwwx reads:
> #1234     1727  0.0  0.1  2132 1000 pts/4    S    22:02   0:00             
\_ /bin/sh -c ../../bin/R --vanilla < tcltk-Ex.R > tcltk-Ex.Rout
2>&1 || (mv tcltk-Ex.Rout tcltk-Ex.Rout.fail && exit 1)
> #1234     1728  0.1  2.6 24524 17004 pts/4   S    22:02   0:03             
\_ /tmp/buildd/r-base-1.6.2.cvs.20030403/bin/R.bin --vanilla
> #1234     1731  0.0  0.0     0    0 pts/4    Z    22:02   0:00             
\_ [R.bin] <defunct>
> 
> strace -p 1728 doesn't show anything useful, but top shows no CPU
> being consumed by the process (which is strange).
Hm, I have not seen that. But Doug Bates mentioned something about building
remotely, and having had a problem.

I have never seen this, and am building both locally and remotely.
> In case this makes any difference, I am redirecting the stdout and
> stderr of pbuilder into a pipe, while leaving stdin unchanged.
Don't know, it may. But that does not propagate to what happens inside
pbuilder, i.e. in the processes started by it, does it? 

Most of the other build systems work, see
     http://buildd.debian.org/build.php?pkg=r-base
The most recent hppa error looks spurious and related to Build-Depends.

Just to confirm, you are building on m68k, correct?  

Dirk


-- 
Wishful thinking can dominate much of the work of a profession for a decade,
but not indefinitely.   -- Robert Shiller, on Efficient Markets models, 2002