fkiss-x, aka Yav's FKiSS viewer for X-Windows (latest version: 0.33a)
(fkiss033.tgz, fkiss033a.tgz)

By the inventor of FKiSS2. Found at the KiSSin Institute of Softwear, which has versions 0.25 (originally: fkiss-x.tar.gz) to 0.33 and the patch for the latest version, 0.33a. These are sources, and have to be compiled. They include a KiSS man page; Linux users will know what that is. The viewer runs from the console in X-Windows and has many command-line options - maximum number of alarms, whether to allow FKiSS or not, etc. - which are explained in the man page and can be preset in the .fkissrc file in the user's home directory. As the default maximum number of alarms for this viewer is 256, this file requires some editing to make sure the viewer doesn't exit while loading a set. One setting is "-geom" to set the window size, although, depending on the window manager, the window can also be resized by dragging, in which case objects dragged off the playfield don't disappear under the playfield's border, but visibly lie on top of it.

Archive support: internal
FKiSS: 2 (2.1 as of version 0.30)
CKiSS: yes
Enhanced palette: no
Bugs and idiosyncrasies:

See the history in the included README file. The latest patch 33a is to correct the fact that the viewer doesn't always erase its temporary kiss archived compression directories.

Bugs in version 0.33a:
-version() counts in a different way (version 2 = FKiSS2.1).
-Not really a bug, but due to the different way windows work in X-Windows, it's best not to use alarms under initialize(); they may have expired by the time the set opens.
-Although the viewer is one of the few to implement viewport() and windowsize(), it has no maximum window size, so if the window manager allows it, the window can be expanded beyond the playfield. A fixed maximum size lets the viewer resize the window back to normal from any size using the maximum dimensions as arguments, but that won't work in this viewer, so the windowsize()/viewport() test set produces very strange results.
-The lock value of an object is the lock values of the cels added up, and a value above 32766 produces an error message. The adding of values is simply non-standard; the bug is that only values above 32767 should produce this error message. Possibly as a result of this, objects with a lock value of 32767 are unfixed.
-setfix() can unfix a locked object but can't lock an unfixed object, ie. it can't set a fix value higher than 0.

In an early version (0.25) the author implemented four non-standard FKiSS2 events: upside(), downside(), leftside() and rightside(). I don't know if they work in the latest version.

Click here for screenshots.