initial white gap/space/band at top of window (remote VNC or XWindows issue?)
Opened this issue · 1 comments
I've recently noticed that Ferret (not PyFerret) has been sizing its windows incorrectly, when using TigerVNC to remotely access my GFDL workstation. The Ferret windows are initially 9.5% too tall, with a white gap at the top. Do you reproduce this?
To demonstrate:
NOAA/PMEL TMAP
FERRET v7.52 (optimized)
Linux 2.6.32-754.23.1.el6.x86_64 64-bit - 12/03/19
15-May-20 15:19
yes? set win/size=.5/asp=1
yes? use coads_climatology
yes? go magnify
yes? shade/l=1 sst
yes? frame/file=plot.gif
yes? spawn file plot.gif
plot.gif: GIF image data, version 87a, 682 x 747
To see the white gap, you might need to actually click on the image to bring it up in its own browser window, so that it appears against a black background, instead of the white background of GitHub. Note that the Unix "file" command indicates that despite the ASPECT=1 qualifier, the output GIF is 682x747, i.e. 9.5% too tall.
There seems to be an initial disconnect between the displayed window and the plot inside it. Ferret appears to be unaware that this gap region exists. If I try to apply a label that extends vertically into the gap, the label is cut off exactly at the edge of the gap. And if I grab the top edge of the window and drag it down to where it's supposed to be (at the top edge of the plot), then everything looks fine; no more gap, and the plot doesn't initially shrink. Only if I keep dragging the top edge of the window down across the actual plot, does the plot then start to shrink (as expected) to fit into that narrower vertical window, thereby creating a new gap to the right of the plot. So the plot seems fine -- it's just that the initial window height isn't being set properly. This issue does not affect the metafile/PostScript output; only the displayed graphic and the output GIF. And it doesn't occur in PyFerret.
This does not appear to be new; it shows up in versions all the way back to Ferret v6.8.2 from August 2012. Perhaps this has been happening all along, and I just noticed it due to so much telework? Or maybe there was a recent update to XWindows or other software at GFDL, that changed Ferret behavior? Not sure.
I'm not seeing it. I connect using Pulse Secure VPN, and X-win32 for X-windows.
How does the saved plot compare with a plot made when starting Ferret via "ferret -gif" ? When I do this the plots are identical. We haven't seen a change in benchmark outputs over time; the tests include plots made both in interactive mode and in -gif mode.