A few comments:
buildconf.inc -- Is there a reason OPT_REMOTE_SYNCH is defined as 0 instead of 1? I think CHDK mostly uses #if instead of #ifdef for these, but this patch uses #ifdefs, causing 0 to enable the code just like an 1 would, which is sort of confusing.
It should be noted that this define does not effect code size (with any significance, that is), it just hides the menu items. In that sense it's different from other entries currently in buildconf.inc.
The patch renames config 257 from "Enable Remote" to "Enable scriptless Remote". Is that correct? The way I remember it, you need to have config 257 enabled for remote to work in scripts (note: we shouldn't force remote enable in scripts, because PTP can be made work in REC mode but remote enable screws that up).
I don't use any icons, I prefer text for everything. That's why I don't really have much of an opinion on how they look as long as they take as little RAM in the binary as possible when disabled. Especially the space icon drawing code seems quite complex (both before and after this patch). Has anyone tried or estimated whether storing and drawing these icons as bitmaps would speed CHDK up or save RAM, or the opposite? Loading icons from the SD card whenever enabled (like menu icon font) would have the good effect that then disabling them saves RAM without rebuilding with the code removed. It may not be much now that there aren't many icons, but if RAM usage of a new vanity/usability feature is optional, it's definitely easier to accept.
Line 313 of patch has tabs, surrounding code doesn't?
USB indicator is a good addition. I'd rather have a text "USB" instead of an icon, though. But I don't use remotes a lot so it doesn't matter to me much.