EUMETCastView on GNU/Linux
Ernst Lobsiger
Dear all
Here is my HOWTO version 0.2. Meanwhile Christian Peters has managed to do what I describe. I hope my effort and paper proves useful for others as well. Many thanks to Hugo for his support. Regards, Ernst P.S. If you have success please report it here indicating PC specs, distribution and graphics setup.
|
|
mkempfde
Hi Ernst, I had no problems following your Howto and successfully compiled and installed on a 4y old PC Intel i3 with 8GB RAM running Ubuntu 16.03LTS 64bit with a NVIDIA GT1030 graphics card (proprietary NVIDIA drivers).
markus@galilei:~/EMCV$ file EUMETCastView runs perfectly smooth and fast, no visible GL errors with the
globe. Am 13.02.2018 um 18:06 schrieb Ernst
Lobsiger:
Dear all
|
|
Christian Peters
Ernst and all,
toggle quoted messageShow quoted text
just compiled EMCV 1.3.3 on Linux with no problems, running smooth. My specs (lswh -short) system MS-7502 bus MS-7502 memory 128KiB BIOS processor Intel(R) Core(TM)2 Quad CPU Q9400 @ 2.66GHz memory 32KiB L1 cache memory 1536KiB L2 cache memory 8GiB System Memory display GK208 [GeForce GT 710B] (MSI NVIDIA GEFORCE GT710) Linux version 4.9.0-4-amd64 (debian-kernel@...) (gcc version 6.3.0 20170516 (Debian 6.3.0-18) ) #1 SMP Debian 4.9.65-3+deb9u1 (2017-12-23) NVIDIA driver: NVIDIA-Linux-x86_64-390.25 Regards, Christian Am 13.02.18 um 18:06 schrieb Ernst
Lobsiger:
Dear all
|
|
Ernst Lobsiger
Hugo and Christian
I still explore its usage but it seems to me EUMETCastView 1.3.3 misplaces OLCI image labels (?) : I load segments select 1 segment OLCI EFR on the la/lon map press make image view image, it has no label top left if I select the "Geostationary" tab in the toolbox I get a black image with yellow label "OLCIEFR' top left The same procedure applies to OLCIERR @Christian: can you reproduce that under GNU/Linux or Windows? Regards, Ernst
|
|
Christian Peters
Ernst,
toggle quoted messageShow quoted text
just tested on Win, same here. No label at the EFR image, black image an "OLCIEFR" on black background. Regards, Christian Am 19.02.18 um 21:45 schrieb Ernst
Lobsiger:
Hugo and Christian
|
|
Hugo
Ernst and Christian,
I tested it and indeed something went wrong the labeling. I will try to solve this for the next release ... grts, Hugi
|
|
Ernst Lobsiger
Hugo
Here is a couple of problems I face. The overlay question might be a feature not a bug. But segfaults in 1.3.3 and now even more in 1.3.4. See image and text attached. Regards, Ernst
|
|
Ernst Lobsiger
Hugo and all
I am wrestling with my graphical environment. Screen shot way too big (sorry) and my text file insists to be a binary (?). Here is exactly the same file renamed Segfaults.txt. Regards, Ernst
|
|
Hugo
Ernst,
I can reproduce the bugs in EUMETCastView. I will try to solve this a.s.a.p. grts, Hugo
|
|
Hugo
Ernst,
I solved the bugs. (I hope) You can clone the source from github. The "modus operandi" is as follows for selecting a new image from the geostationary sats : You can switch tabs (other geo sat) and the last image and overlay will remain visible. Once you select a new row from the list of any of the geo sats, than the image will be cleared and a new overlay will be calculated. The size will not be the correct one, because this is only changed once you make an image by selecting a spectral band. grts, Hugo
|
|
Ernst Lobsiger
Hugo
Thanks for the fix of the overlay and FY 1km + CLAHE segfault. I still have segfaults when I try to image Electro L2, GOES 15 or Himawari 8. It worked in 1.3.3. I copied the *.ini files, is that a problem? Regards, Ernst (... going to bed now)
|
|
Hugo
Ernst,
I uploaded a change in the source code to github for these segfaults. ( Some interference with my new RGB recipes code ...) Everything seems to work now in Linux. grts, Hugo
|
|
Ernst Lobsiger
Hugo,
Many thanks. I noticed you added -no-pie and I get clickable apps now. You may want to add "Meteosat-11" to the "About" (that's all I found :-). I am making my first real steps in git now. "Pro Git" has 508 pages... I already noticed that git preserves the +x bits I am talking about in my HOWTO. So if you strictly commit/upload from the GNU/Linux side this should not happen. But once these +x bits are set (most probably when uploading from an ntfs volume on Windows, or simply when uploading files that traversed somehow a Windows volume?) they will stay on github.com unless you change that and commit it again in GNU/Linux. That's at least my understanding so far ... Regards, Ernst
|
|
Hugo
Ernst,
Yes probably I did a push once from a Windows machine ... I did a complete reset of all the file permissions on github. I downloaded a zip file and it seems ok now. grts, Hugo
|
|
Ernst Lobsiger
Hugo,
Thank you for cleaning this up. It will shorten my next HOWTO version :-). @All: I am working on my GNU/Linux receiver side again. Could somebody post an EumetcastWatcher.ini that manages all possible satellite products that can be treated with the current version of EUMETCastView? Thanks! Regards, Ernst
|
|
Hugo
Hi Ernst,
this is the one I use for the moment ... grts Hugo
|
|
Ernst Lobsiger
Hi Hugo,
Thank you for your EumetCastWatcher.ini. I improved my command line watcher program and tuned my receiver to NFS provide all the data EUMETCastView can handle under the "EUMETCast" tab. Now I'm back on the graphic processing PC and played some hours with you latest version of EUMETCastView. Attached my findings. Maybe some of them could be considered in a next version. I look forward to the future RGB possibilities that you are working on. Regards, Ernst
|
|
Hugo
Ernst,
Thank you for your tests on the program. It is indeed hard to predict every click with the mouse a user will make ... I already solved all the segfaults in the program.( source is now on github ). For the image labels I propose to eliminate this, and put all the image data in the bottom of the Toolbox. I forgot to update the segment numbers in the cylindrical map, still have to do that .. A 'rotate' button should be possible for VIIRS . Kind regards, Hugo
|
|
@lumyk
Hi Ernst,
Thanks for nice description. It was easy to follow your instructions and everything went completely smoothly. At the first sight EUMETCastView 1.4.0 runs quickly and without any visible problem - globe shows constantly ~60 paintGL calls/s. Built and launched on my main computer (with plenty of memory): $ cat /etc/lsb-release | grep DESC ; uname -a; lscpu | grep "Model name"; glxinfo | grep OpenGL | head -3 DISTRIB_DESCRIPTION="Linux Mint 18.3 Sylvia" Linux mentos 4.8.0-58-generic #63~16.04.1-Ubuntu SMP Mon Jun 26 18:08:51 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux Model name: Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz OpenGL vendor string: NVIDIA Corporation OpenGL renderer string: GeForce GTX 970/PCIe/SSE2 OpenGL core profile version string: 4.5.0 NVIDIA 384.130
|
|
hb9erg.kd0hop@...
Hi all,
|
|