[Lazarus] Store empty string property in *.lfm

[Hans-Peter Diettrich]

theo schrieb:

> Yes, this is probably cleaner than my workaround with "none".
> But it needs one extra boolean field, and the "none" solution should be 
> safe too.
> (Afaik there is no system where "none" without slashes is a valid 
> absolute path).

I typically use '<none>' to signal invalid entries, easily recognizable 
by the first character. A non-visual solution were e.g. a #0, that 
should be an illegal value in most cases, but still is a non-empty string.

DoDi