Input handling could probably be improved and considered
prioritary. When increasing the refresh rate like with 0 delay
using more CPU, it's more responsive. This should be configurable.
+ Oddly, when using 0, it can result in a type of progressing chaos and
+ even X11 video garbage. Other exampes where application output interferes
+ with input are aafire/cacafire and sometimes even cmatrix (on slow systems
+ like RPI2). The "chaos" seems mitigated when disabling multiple threads.
CONTROLS
at_vt100_printc (analogterm_sim_t *sim, struct terminal_controller_data *state,
unsigned char c)
{
- analogterm_state_t *st=sim->st;
/* XXX Why not simply use constants that allow more efficient literal math?
* Presumably for an eventual dynamic window size or other terminal modes
* like 132 columns...
int i, many;
int start, end;
+ analogterm_state_t *st=sim->st;
+
/* Mostly duplicated in phosphor.c */
+ /*
+ * XXX Now mostly rewritten after starting from apple2-main.c
+ * If readapted to xscreensaver in general, it would be a good idea to
+ * isolate terminal emulation as a library in utils/.
+ * It could then be used by phosphor, a2 and analogterm, perhaps others.
+ * The target emulation is now xterm. See utils/textclient.c
+ * textclient_putc() for other changes related to PC keyboard input,
+ * and textclient_pipe(). And see analogtv.c changes for less blurry text
+ * and color filtering.
+ *
+ * MMondor 2022
+ */
switch (state->escstate)
{
for (y = 0; y < 8; y++) {
pp = &sim->inp->signal[(ANALOGTV_TOP + 3) + (8 * st->cursy) + y]
[(ANALOGTV_PIC_START + 100) + (7 * st->cursx)];
- for (x = 0; x < 8; x++)
+ for (x = 0; x < 7; x++)
*pp++ ^= 64/*TMODE_NORMAL_LEVEL*/;
}
}
Repositories / xscreensaver-5.35.git / commitdiff