search

MichaelChirico / r-bugs

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

[BUGZILLA #3868] make check fails 1.7.1 with R-release patch in regression tests #1846

Closed MichaelChirico closed 4 years ago

MichaelChirico commented 4 years ago

From: khendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> Full_Name: Kevin B. Hendricks Version: 1.7.1 patched with R-release.diff OS: Linux x86-64 Submission from: (NULL) (65.93.132.185)

Hi,

Built from R-1.7.1 source with R-release.diff patch attached on a RedHat 9 system for Linux x86-64 (dual Opteron system) runing in 64 bit mode.

Build completes fine. But "make check" fails (R is killed because Out-Of-Memory) during reg-tests-1.R. I have verified that the R-Release.diff patch has been applied and the changes to the stopping critieria in that file have been made.

My x86_64 machine has over 2 gig of main memory.

Is the stopping cirtieria a function of the size of the machine registers in use?

All other tests seem to pass with no problems. I have reproduced this error condition with and without optimized "atlas" libs in the build and even with -O0 (to remove possiblility of compiler optimization error).

I am stumped on how to proceed. Obviously the stopping criteria for some method is never being reached resulting in the out of memory condition.

Any hints on how to track this down further appreciated. All altas tests run and pass with flying colors (if that helps).

Thanks,

Kevin

METADATA

MichaelChirico commented 4 years ago

From: Dirk Eddelbuettel <edd@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> On Thu, Aug 21, 2003 at 03:22:17PM +0200, khendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> wrote:

Full_Name: Kevin B. Hendricks
Version: 1.7.1 patched with R-release.diff
OS: Linux x86-64
Submission from: (NULL) (65.93.132.185)

Hi,

Built from R-1.7.1 source with R-release.diff patch attached on a 
RedHat 9 system for Linux x86-64 (dual Opteron system) runing in 64 bit mode.

Build completes fine.  But "make check" fails (R is killed because
Out-Of-Memory)
during reg-tests-1.R.   I have verified that the R-Release.diff patch has
been
applied 
and the changes to the stopping critieria in that file have been made. 

My x86_64 machine has over 2 gig of main memory.

What I your compiler? The GNU GSL people recetnly had to alter one of their regression tests as gcc 3.3.1 (if I recall correctly) would all of a sudden consume all available memory.

Just a quick shot in the dark, this may of course an entirely different issue... But if you can try an older compiler, that may be worth the few minutes it takes.

Dirk

-- Those are my principles, and if you don't like them... well, I have others. -- Groucho Marx

METADATA

MichaelChirico commented 4 years ago

From: "Kevin B. Hendricks" <kevin.hendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> Hi,

Please close the bug, the problem was mine. Sorry about that!

It seems I did not recursively change ownership after I unpacked the R-1.7.1 source file and the "tests" directory ended up with some weird user and group id that for some reason made the following line in reg-tests-1.R fail by using up all memory!

((info <- file.info("."))

Once doing the following, all now works fine ...

tar -zxvf R-1.7.1.tgz gunzip R-release.diff.gz chown -R kbhend R-1.7.1/ (missed this step previously) cd R-1.7.1 patch -p1 < ../R-release.diff

Then building and and running make check passes.

So strange ... I am surprised that trying to get info on the current directory (in this case "tests") when not actually owning the directory could create such a problem. There must be a bug someplace in the x86_64 system calls that let's this happen. It should have just come back with an error message saying I had no priviledges to do that.

Oh well, I guess that is the problem when using a new system like x86_64.

Thanks for your fast response. Please close the bug as user error.

Thanks again,

Kevin

On Thursday 21 August 2003 09:39, Dirk Eddelbuettel wrote:

On Thu, Aug 21, 2003 at 03:22:17PM +0200, khendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> wrote:
> Full_Name: Kevin B. Hendricks
> Version: 1.7.1 patched with R-release.diff
> OS: Linux x86-64
> Submission from: (NULL) (65.93.132.185)
>
>
> Hi,
>
> Built from R-1.7.1 source with R-release.diff patch attached on a
> RedHat 9 system for Linux x86-64 (dual Opteron system) runing in 64 bit
> mode.
>
> Build completes fine.  But "make check" fails (R is killed because
> Out-Of-Memory)
> during reg-tests-1.R.   I have verified that the R-Release.diff patch has
> been applied
> and the changes to the stopping critieria in that file have been made.
>
> My x86_64 machine has over 2 gig of main memory.

What I your compiler?  The GNU GSL people recetnly had to alter one of
their regression tests as gcc 3.3.1 (if I recall correctly) would all of a
sudden consume all available memory.

Just a quick shot in the dark, this may of course an entirely different
issue... But if you can try an older compiler, that may be worth the few
minutes it takes.

Dirk

METADATA

MichaelChirico commented 4 years ago

From: Peter Dalgaard BSA <p.dalgaard@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> khendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> writes:

Built from R-1.7.1 source with R-release.diff patch attached on a 
RedHat 9 system for Linux x86-64 (dual Opteron system) runing in 64 bit mode.

Build completes fine.  But "make check" fails (R is killed because
Out-Of-Memory)
during reg-tests-1.R.   I have verified that the R-Release.diff patch has
been
applied 
and the changes to the stopping critieria in that file have been made. 

My x86_64 machine has over 2 gig of main memory. 

Is the stopping cirtieria a function of the size of the machine registers in
use? 

All other tests seem to pass with no problems.  I have reproduced this error
condition
with and without optimized "atlas" libs in the build and even with -O0 (to
remove possiblility of compiler optimization error).

I am stumped on how to proceed.  Obviously the stopping criteria for some
method
is never being
reached resulting in the out of memory condition.

Any hints on how to track this down further appreciated.  All altas tests run
and pass with flying colors (if that helps).

Hmm.

(1) Does it happen without the patches?

(2) The canonical way to investigate is to

    (a) figure out which file is being processed and rerun R with
    output to the terminal. I.e.

    export SRCDIR=../R/tests # must be set, value depends on build setup 
    bin/R --vanilla < $SRCDIR/reg-tests-1.R

    This avoids buffering and so the last few lines should tell
    you approximately where things go wrong.

    You might want to cut the file down a bit, but in this case it
    is probably not necessary.

    (b) Run R under the debugger 
    bin/R -d gdb
    run --vanilla  < $SRCDIR/reg-tests-1.R

    when it hangs, try hitting control-C and do a "bt" (backtrace)
    command in gdb. You can then poke around in the stack and see
    if you can spot the problem. Or get back to us with some new
    clues.

-- O_ ---- Peter Dalgaard Blegdamsvej 3
c/ /' --- Dept. of Biostatistics 2200 Cph. N
<CENSORING FROM DETECTED PHONE NUMBER ONWARDS; SEE BUGZILLA>

METADATA

MichaelChirico commented 4 years ago

From: "Kevin B. Hendricks" <kevin.hendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> Hi Peter,

Thanks for your response.

I tracked it down to (info <-file.info(".")) hanging and then started checking the obvious as to what permissions that "tests" had and soon found out that everything was quite messed up (I had forgot to chown the files after unpacking them).

Once I cleaned that up and changed owership to a valid user id the hangs went away.

So please close this as user stupidiity on my part. I simply forgot that tar keeps old user ids when unpacking and that I need to fix that myself before patching and building.

Thanks gain for your fast responses!

Kevin

On Thursday 21 August 2003 10:30, Peter Dalgaard BSA wrote:

khendricks@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> writes:
> Built from R-1.7.1 source with R-release.diff patch attached on a
> RedHat 9 system for Linux x86-64 (dual Opteron system) runing in 64 bit
> mode.
>
> Build completes fine.  But "make check" fails (R is killed because
> Out-Of-Memory)
> during reg-tests-1.R.   I have verified that the R-Release.diff patch has
> been applied
> and the changes to the stopping critieria in that file have been made.
>
> My x86_64 machine has over 2 gig of main memory.
>
> Is the stopping cirtieria a function of the size of the machine registers
> in use?
>
> All other tests seem to pass with no problems.  I have reproduced this
> error condition
> with and without optimized "atlas" libs in the build and even with -O0
> (to remove possiblility of compiler optimization error).
>
> I am stumped on how to proceed.  Obviously the stopping criteria for some
> method is never being
> reached resulting in the out of memory condition.
>
> Any hints on how to track this down further appreciated.  All altas tests
> run and pass with flying colors (if that helps).

Hmm.

(1) Does it happen *without* the patches?

(2) The canonical way to investigate is to

(a) figure out which file is being processed and rerun R with
output to the terminal. I.e.

export SRCDIR=../R/tests # must be set, value depends on build
setup bin/R --vanilla < $SRCDIR/reg-tests-1.R

This avoids buffering and so the last few lines should tell
you approximately where things go wrong.

You might want to cut the file down a bit, but in this case it
is probably not necessary.

(b) Run R under the debugger
bin/R -d gdb
run --vanilla  < $SRCDIR/reg-tests-1.R

when it hangs, try hitting control-C and do a "bt" (backtrace)
command in gdb. You can then poke around in the stack and see
if you can spot the problem. Or get back to us with some new
clues.

METADATA

MichaelChirico commented 4 years ago

Audit (from Jitterbug): Thu Aug 21 17:53:40 2003 thomas moved from incoming to not-reproducible

METADATA