Created attachment 1950633 [details] Postscript file for testing Description of problem: If I want to view Postscript file with -sDEVICE=display, 'gs' segfaults Version-Release number of selected component (if applicable): ghostscript-10.0.0-3.fc38.x86_64 How reproducible: Always Steps to Reproduce: 1. $ gs -sDEVICE=display <file.ps> Actual results: Segfault Expected results: GUI viewer for Postscript files starts
Created attachment 1950635 [details] Full backtrace
Note that the x11 device is named 'x11'. Has 'display' ever worked here? Not on 9.55 nor 9.56. (gs should fail more gracefully, of course.)
Not sure - I've tested 'display' device just because x11 device is missing since 10.0.0. From upstream regarding x11 device - it's a dynamic device which got removed because some symbols became hidden, so now only the library is shipped and not a dynamic device. Users can use 'gsx' from ghostscript-gtk to view postscript files. According to upstream 'display' is badly chosen name from the past, it is actually an in-memory rendering device used for rendering to a display on MS Win. But still it would be great if it didn't crash :)
Interesting. If I build 10.0.0 from ghostpdl.git (plus extract.git, minus the libs we de-bundle in fedora) I get the x11 device. I guess the distribution tarball (which the spec pulls, as it should) differs. We have gsx in ghostscript-gtk in fedora. And yes, crashing should not be the way in which gs communicates non-working devices or parameters ... But that should be an upstream bug.
This package has changed maintainer in Fedora. Reassigning to the new maintainer of this component.
The display device is used by the MS Windows, OS/2 and the gtk+ versions of Ghostscript. The segmentation fault has been fixed in v. 10.01.0 and higher. New versions report an unrecoverable error message and exit.
I confirm Alex's findings - the current ghostscript on F38 - ghostscript-10.02.1-2.fc38.x86_64 - gets unrecoverable error instead of segfault, which is correct.