+TODO: underline cursor, simply using underline
+TODO: better default settings for hangul: ex. 냻
+22:32 <zrajm> I *really* like the pointerBlank feature -- though I'd like it to *only*
+ hide the pointer when I hit a key, never on a timeout -- I kinda
+ expected that it would work this way if I set pointerBlankDelay to 0 but
+ that more or less turns the pointer completely off inside the urxvt
+ window.. not the behaviour I expected.
+22:43 <zrajm> E.g. xterm has a couple of settings for when you select using
+ tripple-clicking (entire lines). I've become quite used to
+ "xterm*cutToBeginningOfLine: off" (which causes the first line to be
+ selected only from the beginning of the word under the cursor) and
+ "xterm*cutNewline: off" (which causes xterm not to include the newline
+ character of the last line selected) -- these are especially swell when
+ I select urls (which usually occur on a line by the
+TODO: custom keymap must not "overwrite" default keymap
TODO: do font request caching, might help rxvtd on some machines
TODO: safer command: keymap processing.
TODO: "slow" rendering mode for bidi and scripts
WISH: tabbed windows (hey, just use screen...)
WISH: just for fun, do shade and tint with XRender.
+5.5 Sat Apr 23 22:31:36 CEST 2005
+ - re-enabled modifer state matching as in 5.3, but implement
+ a new builtin: keysym style that simply reverts to the built-in
+ meaning of the given key.
+ - use scrollcolor for next scrollbar instead of hardcoded antiquewhite
+ (patch by Martin Edlman).
+
+5.4 Mon Apr 18 00:33:31 CEST 2005
+ - match modifier state exactly in custom keyboard code.
+ this is necessary because it has no knowledge about
+ built-in mappings and so will overwrite them. the priority-code
+ will be re-enabled when this has changed, so watch out.
+ - cursor blinking now depends on frills.
+ - underline cursor mode (frills, -uc).
+ - implement special value (*g*) for pointerBlankDelay
+ to disble it.
+
5.3 Sun Mar 13 00:20:44 CET 2005
- fix a bug that allowed to overflow a buffer via a long
escape sequence, which is probably exploitable (fix by