Re: [ferret_users] Announcing Ferret v6.6.7, a bug-fix release

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi all,
This was happening only on 64-bit machines. I re-linked the gksm2ps 
program which translates metafiles to postscript.  The incorrect line

   /lw {0.000000 mul 3000 div setlinewidth} def

is now translated to the correct version,

   /lw {3000 div setlinewidth} def

I have replaced the fer_executables tar file for 64-bit, so that now it 
contains the updated gksm2ps executable, and it can be downloaded from 
the 64-bit downloads page,
http://ferret.pmel.noaa.gov/static/Downloads/linux_x86_64_nc4_downloads.html

and also the source code file, with a change to the Makefile for gksm2ps.

Ansley

On 2/16/2011 10:07 AM, Ansley Manke wrote:
> Hi Marco and Jian,
> Thanks for the further information. I'll check into it.
>
> Ansley
>
> On 2/16/2011 9:21 AM, Marco Steinacher wrote:
> > Hi,
> >
> > I see a similar issue with Ferret 6.6.7. Apparently the linewidth
> > definition in the PS files generated by Fprint/gksm2ps has changed. In
> > earlier versions the PS files usually contained the following line:
> >
> > /lw {3000 div setlinewidth} def
> >
> > Now it looks like this:
> >
> > /lw {0.000000 mul 3000 div setlinewidth} def
> >
> >
> > I used to increase the line width of the PS output with ps_thicken as
> > described in the FAQ
> > (http://ferret.pmel.noaa.gov/Ferret/faq/changing-line-thickness-in-postscript-files). 
> >
> > It uses sed to change the PS file. E.g.
> >
> > sed -e "s/^\/lw {\(.*\) div setlinewidth/\/lw {4 mul \1 div 
> > setlinewidth/"
> >
> > This does not work anymore. Now, one would have to do something like
> >
> > sed -e "s/^\/lw {0.000000 mul \(.*\) div setlinewidth/\/lw {4 mul \1 div
> > setlinewidth/"
> >
> >
> > Marco
> >
> >
> > Jian Ma wrote:
> > > Hi Ansley and All,
> > >
> > > I found a bug after upgrading from 6.5 to 6.6.7. When I transfer my 
> > > plot
> > > to eps with gksm2ps, the thickness of the lines are 1, no matter 
> > > what is
> > > set in the jnl. One of my scripts works fine with the former version 
> > > but
> > > has this problem after the upgrade.
> > >
> > > I did not delete my older version during the upgrade, though, to keep
> > > some customized scripts. Would this be the cause? Anybody has
> > > experienced similar problem?
> > >
> > > Thanks,
> > > Tony
> > >
> > > On Mon, 2011-02-14 at 10:07 -0800, Ansley Manke wrote:
> > > > Hi Martin,
> > > > We've just switched over to RHEL5 and libc.so.6 on our machine 
> > > > shows as
> > > > linking to libc-2.5.so.  (Previous to that switch we were hearing from
> > > > users with newer linux systems who had library mis-matches, though
> > > > compatibility libraries generally lets one find the right library for
> > > > code built on an earlier operating system to run on a newer one.)
> > > >
> > > > The executable in the tar file is already linked with all of the 
> > > > static
> > > > libraries we had available to us.  We're very much aware  of this 
> > > > issue,
> > > > and are thinking it may be time to work towards releasing source code
> > > > with a nice build process so people could build Ferret themselves. 
> > > > Even
> > > > as things stand now, building from source is not difficult, but does
> > > > require building HDF and NetCDF libraries. I haven't posted the source
> > > > code for this fixed version, but I'll do that later today.
> > > >
> > > > We'll continue to work on a good general solution for this problem.
> > > >
> > > > Ansley
> > > >
> > > >
> > > >
> > > > On 2/14/2011 5:15 AM, Martin Schmidt wrote:
> > > > > Dear Ansley,
> > > > > I tried to install Version 6.6.7 at SLES10 but failed. 
> > > > > libselinux.so.1
> > > > > is missing. I tried to use a the lib from SLES11.
> > > > > Now I end up with
> > > > >
> > > > > ferret_v667: /lib64/libc.so.6: version `GLIBC_2.8' not found 
> > > > > (required
> > > > > by /linsoft/ferret_v667_64_dynamic/lib/libselinux.so.1)
> > > > >
> > > > > Indeed, my libc.so.6 is a link to libc-2.4.so with SLES10. With 
> > > > > SLES11
> > > > > the link is to libc-2.11.1.so.
> > > > >
> > > > > Is there any chance to benefit from the latest bug fixes at machines
> > > > > that are not using the latest system version. Possibly by a more 
> > > > > complete
> > > > > static linking?
> > > > > We are not able to update the system, because the system 
> > > > > management is
> > > > > linked with the hardware contract.
> > > > >
> > > > > Greetings,
> > > > > Martin Schmidt
> > > > >
> > > > > Ansley Manke wrote:
> > > > > > Hi all,
> > > > > > There is a new release of Ferret available now, Version 6.6.7, which
> > > > > > corrects a recently reported bug in Ferret v6.6.5.
> > > > > >
> > > > > > In Ferret v6.6.5, the last official Ferret release, the
> > > > > > multi-dimensional computation of a definite integral @DIN over X and
> > > > > > Z, but not integrating over latitude, is incorrect. This means that
> > > > > > the computation of quantities such as a meridional transport at a
> > > > > > fixed latitude, will be incorrect when calculated by Ferret v6.6.5.
> > > > > >
> > > > > > The calculation was done correctly in the previous release, Ferret
> > > > > > v6.4, and is corrected in v6.6.7, which may be downloaded from the
> > > > > > Ferret Downloads page at
> > > > > >
> > > > > > http://ferret.pmel.noaa.gov/Ferret/downloads
> > > > > >
> > > > > > The Downloads documentation is updated and the main Ferret
> > > > > > documentation will be updated to reflect this new version number
> > > > > > shortly.
> > > > > >
> > > > > > Ansley