Re: Empty PS/EPS

From: Walter Kolczynski <wckolczy_at_nyahnyahspammersnyahnyah>
Date: Thu Jan 10 2013 - 03:51:46 MST
Empty here means zero length. Dataset independent (running same dataset once sometimes works, sometimes fails).

I believe the GKS errors occur whenever I try to use RasterFill and it produces 0-sized files, but using a different fill does not produce GKS but may still produce 0-sized files.

There's already a testcase on the ftp server I uploaded for Mary. I've updated the script since (and ironed out some bugs of my own creation), but the essence is still there. I can provide the updated script if you'd like.

- Walter


On 09-Jan-13 08:04, Rick Brownrigg wrote:
Hi Walter,

I wondered about this issue when it first appeared and then inexplicably disappeared -- must be sun spots, or something :-/
When you say "empty" PS files, do you mean zero-lengthed files, or empty of graphics (that is, is there at least some postscript preamble written to the files but no actual graphics, or are they just plain empty)?  Is it dataset dependent; i.e., if it works for one dataset, it always works?  And when it fails, do you always get the GKS error cited below?  

I'm not sure what to advise at this point -- If there's anyway to get ahold of a test case that exhibits the problem, that would be hugely useful.

Rick

On Jan 9, 2013, at 5:26 AM, Walter Kolczynski <wckolczy@alum.wustl.edu> wrote:

I've also been having this problem intermittently. Sometimes the PS is created properly, then I'll get only empty files using the same script later. Then it'll work again the next day once or twice before breaking again. PDF works reliably. I believe the problem is related to the following one I posted on the ncl-install list (http://www.ncl.ucar.edu/Support/install_archives/2012/0210.html), which may help isolating the bug. When I set cnFillMode = "RasterFill", I get the following warnings:

cro_Text: can only retrieve current color for solid patterns
warning:GKS:GCA: --GKS NOT IN PROPER STATE: GKS SHALL BE IN STATE GKCL

As requested earlier in this thread, here is the ldd (cairo-trace does not work):

        linux-vdso.so.1 =>  (0x00007fff18bff000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fe5ad360000)
        libbz2.so.1 => /lib64/libbz2.so.1 (0x00007fe5ad151000)
        libXrender.so.1 => /usr/lib64/libXrender.so.1 (0x00007fe5acf47000)
        libX11.so.6 => /usr/lib64/libX11.so.6 (0x00007fe5acc09000)
        libXext.so.6 => /usr/lib64/libXext.so.6 (0x00007fe5ac9f7000)
        libm.so.6 => /lib64/libm.so.6 (0x00007fe5ac7a1000)
        libdl.so.2 => /lib64/libdl.so.2 (0x00007fe5ac59c000)
        libstdc++.so.6 => /usr/lib64/libstdc++.so.6 (0x00007fe5ac291000)
        libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007fe5ac07a000)
        libc.so.6 => /lib64/libc.so.6 (0x00007fe5abd1b000)
        /lib64/ld-linux-x86-64.so.2 (0x00007fe5ad5aa000)
        libxcb-xlib.so.0 => /usr/lib64/libxcb-xlib.so.0 (0x00007fe5abb19000)
        libxcb.so.1 => /usr/lib64/libxcb.so.1 (0x00007fe5ab8fd000)
        libXau.so.6 => /usr/lib64/libXau.so.6 (0x00007fe5ab6f8000)

[Performed when postscripts were malfunctioning.]

- Walter


On 20-Dec-12 11:57, Carl Schreck wrote:
Not sure what changed, but I think they rebooted the server and now it works fine. If it happens again, I'll run those diagnostics.

Thanks!

   Carl


On Thu, Dec 20, 2012 at 12:26 PM, Rick Brownrigg <brownrig@ucar.edu> wrote:
Hi Carl,

Offhand, I don't know what could cause that.  On the machine where this does not work, I'd be curious as to the output of:

  ldd `which ncl`

and 

  cairo-trace --no-file ncl  yourScript.ncl

The later may not work depending on how things are linked and of course whether cairo-trace is even installed on your machine, but…



On Dec 20, 2012, at 9:56 AM, Carl Schreck <carl@cicsnc.org> wrote:

I'm testing out a new machine here, and I'm having a strange issue. If I output to an PDF or PNG, everything is fine. But if I try to output to EPS or PS, it just makes an empty file.

The same script will do it on one machine, but not the other. And I do delete(wks) at the end. So it shouldn't be the script.

I'm using ncl_ncarg-6.1.0.Linux_RedHat_x86_64_gcc444.tar.gz

The machine that works has: gcc (GCC) 4.5.3
The machine that doesn't has: gcc (GCC) 4.4.6 20120305 (Red Hat 4.4.6-4)

Any ideas?

Thanks and Happy Holidays!

   Carl

--
Carl J. Schreck III, PhD
Cooperative Institute for Climate and Satellites (CICS-NC)
North Carolina State University   |   NOAA's National Climatic Data Center
Email: carl@cicsnc.org   |   Tel: 828-257-3140
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk




--
Carl J. Schreck III, PhD
Cooperative Institute for Climate and Satellites (CICS-NC)
North Carolina State University   |   NOAA's National Climatic Data Center
Email: carl@cicsnc.org   |   Tel: 828-257-3140


_______________________________________________
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
Received on Thu Jan 10 04:45:29 2013

This archive was generated by hypermail 2.1.8 : Mon Jan 14 2013 - 14:31:38 MST