[Rd] xerbla called from BLAS routine (PR#8100)

[Prof Brian Ripley]

On Mon, 29 Aug 2005, Göran Broström wrote:

> On Mon, Aug 29, 2005 at 10:06:00AM +0200, ripley at stats.ox.ac.uk wrote:
>>   This message is in MIME format.  The first part should be readable text,
>>   while the remaining parts are likely unreadable without MIME-aware tools.
>>
>> --27464147-1510948051-1125298763=:26166
>> Content-Type: TEXT/PLAIN; charset=iso-8859-1; format=flowed
>> Content-Transfer-Encoding: QUOTED-PRINTABLE
>>
>> I don't believe your message actually came from R's XERBLA.  Note that the=
>> =20
>> name is not quoted and that the name of the routine is not 6 characters=20
>> long.
>
> Actually, it did.

You seem to be agreeing that it did not (which is the truth).

> I didn't quote the message exactly (as I said).

You didn't say, and the posting guide does ask you to do so exactly.

> Now, I took the liberty to make my own suggested change in 
> R-devel_2005-08-28 and rebuilt, and you may guess what error message I 
> got.;)
>
> Error in coxreg.fit(X, Y, rs, strats, offset, init, max.survs, method,  :
>        The BLAS/LAPACK routine 'DSYR  ' gave error code -7
>
>
> BTW, the original BLAS xerbla (from the netlib page you referred to)
> gives the message:
>
>     WRITE( *, FMT = 9999 )SRNAME, INFO
> 9999 FORMAT( ' ** On entry to ', A6, ' parameter number ', I2, ' had ',
>     $      'an illegal value' )
>
> which I think is even better than my original suggestion.
> (Why was it changed?)

Probably because R is not a Fortran program and so needs to make clear 
what these refer to, and because LAPACK does have other error values that 
invalid parameter values.

>
> Göran
>
>>
>> Here is a genuine example from R:
>>
>> Error in La.svd(x, nu, nv) : LAPACK routine 'DGEBRD' gave error code -10
>>
>> There is no R 2.2.0, BTW.  If referring to R-devel, please do so with an=20
>> SVN revision.
>>
>> On Sun, 28 Aug 2005 gb at stat.umu.se wrote:
>>
>>> Full_Name: G=F6ran Brostr=F6m
>>> Version: R-2.1.1, 2.2.0
>>> OS: Debian unstable
>>> Submission from: (NULL) (213.65.9.59)
>>>
>>>
>>> Some BLAS routines call xerbla for error messages, which results in a mes=
>> sage
>>> like
>>> "LAPACK routine DGER gave error code -9".
>>>
>>> Suggested solution: In
>>>
>>> void F77_NAME(xerbla)(char *srname, int *info)
>>> {
>>>   /* srname is not null-terminated.  It should be 6 characters. */
>>>    char buf[7];
>>>    strncpy(buf, srname, 6);
>>>    buf[6] =3D '\0';
>>>    error(_("LAPACK routine '%6s' gave error code %d"), buf, -(*info));
>>> }
>>>
>>> change 'LAPACK' to 'The' (or 'The BLAS/LAPACK').
>>>
>>> ______________________________________________
>>> R-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>>
>>>
>>
>> --=20
>> 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
>> --27464147-1510948051-1125298763=:26166--
>>
>> ______________________________________________
>> R-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>
> -- 
> Göran Broström                    tel: +46 90 786 5223
> Professor and Head
> Department of Statistics          fax: +46 90 786 6614
> Umeå University                   http://www.stat.umu.se/~goran.brostrom/
> SE-90187 Umeå, Sweden             e-mail: gb at stat.umu.se
>
>

-- 
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