Here's this fellow's compiler info. I don't mean to dump this task onto
you, but I'm not versed in our autobuild system. Its never gone well if I
try to produced a binary for someone else on my machine.
---------- Forwarded message ----------
From: Philippe Papin <ppapin@albany.edu>
Date: Sun, Jun 22, 2014 at 9:43 PM
Subject: Re: Strange pattern fill issue with stippling when
plotting wind barbs.
To: Rick Brownrigg <brownrig@ucar.edu>
Hi Rick,
For some reason I have this email thread as "read" on my gmail account, so
my apologies for being slow to get back to you!.
*gcc -v results in:*
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi --disable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --with-cpu=generic
--host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.2 20080704 (Red Hat 4.1.2-54)
*gfortran -v results in (should be identical to above):*
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi --disable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --with-cpu=generic
--host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.2 20080704 (Red Hat 4.1.2-54)
Let me know if you need any more information. Thanks.
-Philippe
On Fri, Jun 20, 2014 at 4:15 PM, Rick Brownrigg <brownrig@ucar.edu> wrote:
> Hi Philippe,
>
> This fix has been available, and we finally have the cycles available to
> make a custom binary (crazy week!). I wasn't aware of this, but in addition
> to the sysinfo you provided, we also need to know the compiler versions:
>
> gcc -v
> gfortran -v
>
>
> Thanks!
> Rick
>
>
>
> On Fri, Jun 13, 2014 at 8:28 AM, Philippe Papin <ppapin@albany.edu> wrote:
>
>> Hi Rick,
>>
>> Thanks for taking the time to look at this and come up with a fix! My
>> system information is below.
>>
>> Linux ash.atmos.albany.edu 2.6.18-348.16.1.el5 #1 SMP Wed Aug 21
>> 04:00:25 EDT 2013 x86_64 x86_64 x86_64 GNU/Linux
>>
>> Really appreciate the help!
>>
>> -Philippe
>>
>>
>>
>>
>>
>>
>> On Wed, Jun 11, 2014 at 5:32 PM, Rick Brownrigg <brownrig@ucar.edu>
>> wrote:
>>
>>> Hi Philippe,
>>>
>>> After looking through the wind-barb code, I am all but certain your
>>> speculation is correct! It will take some more time to gen-up an example
>>> and test the fix, but the fix looks to be a one-liner. Unfortunately, I
>>> can't come up with a work-around -- this happens at such a low level in the
>>> graphics code, its hard to undo the change-of-state from the NCL level.
>>>
>>> The best I may be able to do is to get you a new ncl binary once a fix
>>> is tested/verified. For that, I would need to know what system you are
>>> using:
>>>
>>> uname -a
>>>
>>> In the meantime, I'm filing this as a bug, and I will also make or amend
>>> a ticket for an option to thicken up patterned lines.
>>>
>>> Thanks for your awesome clues in pinpointing the issue!
>>> Rick
>>>
>>>
>>>
>>> On Sun, Jun 8, 2014 at 9:43 PM, Philippe Papin <ppapin@albany.edu>
>>> wrote:
>>>
>>>> Hi NCL talk,
>>>>
>>>> I've been making a series of images that use the stipple pattern fill
>>>> option for a contour interval. However, I've encountered some strange
>>>> behavior when this stippling is used for multiple images created using a do
>>>> loop. I am using *NCL* *v.6.1.2* to make these images.
>>>>
>>>> The datasets and resources used to create these two figures are
>>>> identical, the only difference being that a counter was incremented using a
>>>> do loop to make another image.
>>>>
>>>> Here are two images:
>>>>
>>>> *First image: *
>>>> http://www.atmos.albany.edu/student/ppapin/lb13_img/test/pure_gyre_sr_comp_850geo_test_0.png
>>>>
>>>> *Second image (after incrementing looper by one)*
>>>>
>>>> http://www.atmos.albany.edu/student/ppapin/lb13_img/test/pure_gyre_sr_comp_850geo_test_1.png
>>>>
>>>> As you can see, the stippling does not display the same way, with the
>>>> second image having "hollow" stipple dots rather than the filled stipple
>>>> dots of the first image.
>>>>
>>>> Here are the resources used to create the pattern fill stippling:
>>>>
>>>> * cssres@cnLinesOn = True*
>>>> * cssres@cnFillOn = True*
>>>> * cssres@cnLevelSelectionMode = "ExplicitLevels"*
>>>> * cssres@cnLevels = (/95/) ; set the contour levels*
>>>> * cssres@cnFillColors = (/"transparent","gray20"/)*
>>>> * cssres@cnLineLabelPlacementMode = "Constant"*
>>>> * cssres@cnLineThicknessF = 3.0 *
>>>>
>>>> * cssres@cnFillPattern = 17*
>>>> * cssres@cnFillScaleF = .7*
>>>> * cssres@cnFillDotSizeF = .0045*
>>>> * cssres@cnLineColor = "gray20"*
>>>> * cssres@cnLineLabelsOn = False*
>>>> * cssres@cnInfoLabelOn = False *
>>>> * cssres@lbLabelBarOn = False*
>>>>
>>>> And when plotting:
>>>>
>>>> *do i=0,1*
>>>>
>>>>
>>>> *mpid = gsn_csm_map(wks,mpres)*
>>>> *vcid = gsn_csm_vector(wks,u850_anom,v850_anom,vcres)*
>>>> *ssid = gsn_csm_contour(wks,g850_ss,cssres)*
>>>>
>>>> *overlay(mpid,ssid)*
>>>> *overlay(mpid,vcid)*
>>>>
>>>> *....*
>>>>
>>>> *end do*
>>>>
>>>> My hypothesis after a little bit of testing is that the vector plot
>>>> option *vcres@vcGlyphStyle = "WindBarb"* is modifying the way the
>>>> stippling is displayed, transforming the stipple dot objects from solid to
>>>> hollow (so that the calm hollow circle for wind barbs can be plotted). The
>>>> problem is these dots don't change back to solid for the *cssres*
>>>> resource the next time the image is plotted.
>>>>
>>>> Is there away around this? I can just rearrange the way I am overlaying
>>>> the patterned contour so that the two images plot the same way, but I'd
>>>> rather have solid filled stippling dots, and that doesn't seem possible
>>>> when I am also plotting wind barb vector objects.
>>>>
>>>> I'd be happy to provide more code to troubleshoot, although the script
>>>> this is from is a bit cumbersome, so I hope these snippets are enough to
>>>> diagnose my problem.
>>>>
>>>> Thanks,
>>>> -Philippe
>>>>
>>>> P.S. I dug around the ncl-talk archives, and saw a previous thread
>>>> stating there currently is no way to make *thicker* pattern lines from
>>>> the available options
>>>> <https://www.ncl.ucar.edu/Document/Graphics/Images/fillpatterns.png> other
>>>> than just changing the scaling. Count me as one of those that would like to
>>>> see an option to make thicker patterned lines.
>>>>
>>>>
>>>> -------------------------------------
>>>> Philippe P. Papin
>>>> Dept. of Atmospheric and Environmental Sciences
>>>> University at Albany, SUNY
>>>> 1400 Washington Ave, Albany, NY 12222
>>>> http://www.atmos.albany.edu/student/ppapin/
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>>>
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>>
>>
>>
>> --
>> -------------------------------------
>> Philippe P. Papin
>> Dept. of Atmospheric and Environmental Sciences
>> University at Albany, SUNY
>> 1400 Washington Ave, Albany, NY 12222
>> http://www.atmos.albany.edu/student/ppapin/
>>
>
>
--
-------------------------------------
Philippe P. Papin
Dept. of Atmospheric and Environmental Sciences
University at Albany, SUNY
1400 Washington Ave, Albany, NY 12222
http://www.atmos.albany.edu/student/ppapin/
Received on Mon Jun 23 03:40:04 2014
This archive was generated by hypermail 2.1.8 : Wed Jul 23 2014 - 15:33:46 MDT
