[BioC] probelms with hexbin (Solaris, R1.9, BioC 1.4)

[Nicholas Lewin-Koh]

Hi Ariel,
The segmentation fault may not be hexbin, but may be due to something in
your compile of R.
We have a sun here running solaris 9 so I can check with our sysop if I
can compile R-2.0 
and hexbin 1.0.10 using 64 bit. I did it a while back for an older
version on Solaris 8 and
had no problems which is why I suspect it is your whole compilation. Can
you try just using 
plot and polygon and changing the pars and see what happens? If that
works than
the next culprit is the hexbin fortran code, and from there .....

Thanks
Nicholas

> Message: 7
> Date: Fri, 8 Oct 2004 12:01:13 -0400
> From: Ariel Chernomoretz <ariel.chernomoretz at crchul.ulaval.ca>
> Subject: [BioC] probelms with hexbin (Solaris, R1.9, BioC 1.4)
> To: bioconductor at stat.math.ethz.ch
> Message-ID: <47869513-1943-11D9-ABF3-0003939715AE at crchul.ulaval.ca>
> Content-Type: text/plain; charset=US-ASCII; format=flowed
> 
> Hi to all,
> I want to use the hexbin package in order to reduce the size of scatter 
> plot files.
> However, after installing hexbin 1.0.8 I found a very unstable behavior
> For example, from a fresh R session, running twice
>  > example(hexbin)
> produces a segmentation fault.
> 
> In addition even with the very simple example provided in the doc
> the following error messages appears (sometimes):
>  > library(hexbin)
>  > x<-rnorm(20000)
>  > y<-rnorm(20000)
>  > hbin<-hexbin(x,y,bins=40)
> Warning message:
> Lost counts in binning in: hexbin(x, y, xbins=40)
>  > plot(hbin)
> Error in plot.hexbin(hbin) : plot width too small
> 
> I also tried with the developmental version 1.0.10 and got a 
> segmentation fault
> (I am not using a dev version of R)
> 
> Solaris 8
> R 1.9.0  (compiled 64 bits)
> Bioconductor 1.4
> 
> I'll appreciate a lot any suggestion.
> Thanks
> ariel./
> 
> 
> 
> Ariel Chernomoretz, Ph.D.
> Centre de recherche du CHUL
> 2705 Blv Laurier, bloc T-367
> Sainte-Foy, Qc
> G1V 4G2
> (418)-525-4444 ext 46339
> 
> 
> 

> ------------------------------
> 
> Message: 12
> Date: Tue, 12 Oct 2004 19:42:10 +0100
> From: Wolfgang Huber <huber at ebi.ac.uk>
> Subject: Re: [BioC] probelms with hexbin (Solaris, R1.9, BioC 1.4)
> To: Ariel Chernomoretz <ariel.chernomoretz at crchul.ulaval.ca>
> Cc: bioconductor at stat.math.ethz.ch
> Message-ID: <416C2582.3020801 at ebi.ac.uk>
> Content-Type: text/plain; charset=us-ascii; format=flowed
> 
> Hi Ariel,
> 
> does the problem persist if you use R-2.0.0 and the newest hexbin?
> Whithout knowing any of the details - has anyone got hexbin running with 
> 64 bit?
> 
> A slightly different, but related functionalility (false color 
> representations of density estimates) is also provided by the function
> smoothScatter in the package prada (development version), you could try 
> whether you find it useful (see the examples).
> 
> Bw
>   Wolfgang
> 
> 
> Ariel Chernomoretz wrote:
> > Hi to all,
> > I want to use the hexbin package in order to reduce the size of scatter 
> > plot files.
> > However, after installing hexbin 1.0.8 I found a very unstable behavior
> > For example, from a fresh R session, running twice
> >  > example(hexbin)
> > produces a segmentation fault.
> > 
> > In addition even with the very simple example provided in the doc
> > the following error messages appears (sometimes):
> >  > library(hexbin)
> >  > x<-rnorm(20000)
> >  > y<-rnorm(20000)
> >  > hbin<-hexbin(x,y,bins=40)
> > Warning message:
> > Lost counts in binning in: hexbin(x, y, xbins=40)
> >  > plot(hbin)
> > Error in plot.hexbin(hbin) : plot width too small
> > 
> > I also tried with the developmental version 1.0.10 and got a 
> > segmentation fault
> > (I am not using a dev version of R)
> > 
> > Solaris 8
> > R 1.9.0  (compiled 64 bits)
> > Bioconductor 1.4
> > 
> > I'll appreciate a lot any suggestion.
> > Thanks
> > ariel./
> > 
> > 
> > 
> > Ariel Chernomoretz, Ph.D.
> > Centre de recherche du CHUL
> > 2705 Blv Laurier, bloc T-367
> > Sainte-Foy, Qc
> > G1V 4G2
> > (418)-525-4444 ext 46339
> > 
> > _______________________________________________
> > Bioconductor mailing list
> > Bioconductor at stat.math.ethz.ch
> > https://stat.ethz.ch/mailman/listinfo/bioconductor
> 
> 
> -- 
> -------------------------------------
> Wolfgang Huber
> European Bioinformatics Institute
> European Molecular Biology Laboratory
> Cambridge CB10 1SD
> England
> Phone: +44 1223 494642
> Http:  www.dkfz.de/abt0840/whuber
> 
> 
> 
> ------------------------------
> 
> Message: 13
> Date: Tue, 12 Oct 2004 20:19:15 +0100
> From: Wolfgang Huber <huber at ebi.ac.uk>
> Subject: Re: [BioC] probelms with hexbin (Solaris, R1.9, BioC 1.4)
> To: Ariel Chernomoretz <ariel.chernomoretz at crchul.ulaval.ca>,
> 	bioconductor at stat.math.ethz.ch
> Message-ID: <416C2E33.6010802 at ebi.ac.uk>
> Content-Type: text/plain; charset=us-ascii; format=flowed
> 
> 
> Hi Ariel,
> 
> if you care enough, you could compile R with debugging turned on, start 
> a debugger (e.g. "ddd" http://www.gnu.org/software/ddd), start R in 
> there, and see where the crash happens (perhaps it is indeed a problem 
> that has to do with 32bit versus 64bit integers in the C code).
> 
> Bw.
> Wolfgang
> 
> 
> Ariel Chernomoretz wrote:
> > 
> > hi Wolfgang
> > 
> > the problem persists using R-2.0.0 and hexbin 1.0.10. (it's a pitty because
> > hexbin seems to be great and does exactly what i meant to do!)
> > anyway I will give smoothScatter a chance,
> > thanks
> > 
> > ariel./
> > 
> > On 12 oct. 2004, at 14:42, Wolfgang Huber wrote:
> > 
> >> Hi Ariel,
> >>
> >> does the problem persist if you use R-2.0.0 and the newest hexbin?
> >> Whithout knowing any of the details - has anyone got hexbin running 
> >> with 64 bit?
> >>
> >> A slightly different, but related functionalility (false color 
> >> representations of density estimates) is also provided by the function
> >> smoothScatter in the package prada (development version), you could 
> >> try whether you find it useful (see the examples).
> >>
> >> Bw
> >>  Wolfgang
> >>
> >>
> >> Ariel Chernomoretz wrote:
> >>
> >>> Hi to all,
> >>> I want to use the hexbin package in order to reduce the size of 
> >>> scatter plot files.
> >>> However, after installing hexbin 1.0.8 I found a very unstable behavior
> >>> For example, from a fresh R session, running twice
> >>>  > example(hexbin)
> >>> produces a segmentation fault.
> >>> In addition even with the very simple example provided in the doc
> >>> the following error messages appears (sometimes):
> >>>  > library(hexbin)
> >>>  > x<-rnorm(20000)
> >>>  > y<-rnorm(20000)
> >>>  > hbin<-hexbin(x,y,bins=40)
> >>> Warning message:
> >>> Lost counts in binning in: hexbin(x, y, xbins=40)
> >>>  > plot(hbin)
> >>> Error in plot.hexbin(hbin) : plot width too small
> >>> I also tried with the developmental version 1.0.10 and got a 
> >>> segmentation fault
> >>> (I am not using a dev version of R)
> >>> Solaris 8
> >>> R 1.9.0  (compiled 64 bits)
> >>> Bioconductor 1.4
> >>> I'll appreciate a lot any suggestion.
> >>> Thanks
> >>> ariel./
> >>> Ariel Chernomoretz, Ph.D.
> >>> Centre de recherche du CHUL
> >>> 2705 Blv Laurier, bloc T-367
> >>> Sainte-Foy, Qc
> >>> G1V 4G2
> >>> (418)-525-4444 ext 46339
> >>> _______________________________________________
> >>> Bioconductor mailing list
> >>> Bioconductor at stat.math.ethz.ch
> >>> https://stat.ethz.ch/mailman/listinfo/bioconductor
> >>
> >>
> >>
> >> -- 
> >> -------------------------------------
> >> Wolfgang Huber
> >> European Bioinformatics Institute
> >> European Molecular Biology Laboratory
> >> Cambridge CB10 1SD
> >> England
> >> Phone: +44 1223 494642
> >> Http:  www.dkfz.de/abt0840/whuber
> >> -------------------------------------
> >>
> > Ariel Chernomoretz, Ph.D.
> > Centre de recherche du CHUL
> > 2705 Blv Laurier, bloc T-367
> > Sainte-Foy, Qc
> > G1V 4G2
> > (418)-525-4444 ext 46339
> >
> 
> 
>