Re: Compilation of R under Mandrake Linux 6.1 (helios)

About this list Date view Thread view Subject view Author view Other groups

Subject: Re: Compilation of R under Mandrake Linux 6.1 (helios)
From: Duncan Temple Lang (duncan@rice.research.bell-labs.com)
Date: Sat 13 Nov 1999 - 02:48:01 EST


Message-Id: <199911121649.RAA12920@stat.math.ethz.ch>

> Cc: Duncan Temple Lang <duncan@rice.research.bell-labs.com>,
> R-devel@stat.math.ethz.ch
>
> rossini@biostat.washington.edu (A.J. Rossini) writes:
>
> > (ob-reason for posting to R-devel:
> > Has anyone ever seen the error message:
> > "Abort."
> > with R? Thomas hadn't either, and suggested running under
> > gdb, which I hope to, late this afternoon -- it's occuring
> > during a library call, in particular, the R/OMEGAHAT/CORBA
> > package, Duncan :-(...).
>
> I think that happens when the program calls abort(), which seems to
> happen (potentially) two times in memory.c and thrice in saveload.c.
> For the Windows version, there are three cases in regex.c.
>
> Both cases in memory.c happen if one bumps into an object with an
> invalid typeof field during garbage collection. The ones in saveload.c
> all seem to be in the HDF5 code (where it would seem to be a bit of an
> overkill!). The ones in regex seem to be related to inconsistencies in
> the compiled regexp code. My guess would be that there's a memory
> corruption going on, triggering one of the traps in memory.c.
>

It is going to be tricky to diagnose an abort() call without the
actual R command that caused it. My guess would be that there is an
exception being thrown in the Orbacus libraries and that is
propogating its way to abort.

Tony.
>> Sure, but they are two slightly different things. R/GNOME/CORBA is
>> more for standard desktop integration, a bit away from the user; GNOME
>> is pretty useful/stable these days. R/OMEGAHAT/CORBA still seems to
>> need some work, but theoretically will be quite interesting for
>> user-land stuff.

Well the code for integrating R with Gnumeric is not "stable" either.
In fact I am not sure how much, if any, exists at this point. So
comparing the stability Gnome and the R/CORBA mechanism is
irrelevant. Even if there was that connection at the C level, more
would be needed to actual map data from R to CORBA and vice-versa.
This is what the dynamic CORBA facility in R does via "remote references".

Additionally, the R-S/CORBA interfaces are pretty solid here. A few
bug reports to omega-remote@omegahat.org (with perhaps a little more
information than above ;-)) will probably get them fixed pretty
rapidly.

>
> --
> O__ ---- Peter Dalgaard Blegdamsvej 3
> c/ /'_ --- Dept. of Biostatistics 2200 Cph. N
> (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918
> ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
>

-- 
_______________________________________________________________

Duncan Temple Lang duncan@research.bell-labs.com Bell Labs, Lucent Technologies office: (908)582-3217 700 Mountain Avenue, Room 2C-259 fax: (908)582-3340 Murray Hill, NJ 07974-2070 http://cm.bell-labs.com/stat/duncan

"Languages shape the way we think, and determine what we can think about." Benjamin Whorf -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._


About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b25 : Tue 04 Jan 2000 - 14:16:10 EST