--- In [email protected], "Sheri" <sheri...@...> wrote: > > Doc says, if type for default_get_set_type is numeric, it can use values in > the range 0 to 65280. 0x2153 is 8531; but it gives the error: > > unicode.set_chars: Numeric value too big
yup, I was comparing with 0xFF instead of 0xFFFF. Fixed in unicodePlugin0.73_090728.zip in http://tech.groups.yahoo.com/group/power-pro/sfupld?path=/0_TEMP_/AlansPluginProvisional/ > Also, error_dialog_on is documented, but the service doesn't exist. No, I forgot to delete documentation to it. You can have it if yo0u want it, but then I'd have to set unicode_status all the time so scripter would know when things went wrong. Can do, your call. Original idea was to get binary and unicode plugins whose services look as much like bruce's equivalents for standard strings as possible. Playing with unicode_status and error dialog deviates from that, but no reason I shouldn't. > Also I couldn't seem to use ustring if created from using unicode.empty. Ah. Bit of a conceptual pig's ear (pig's conceptual ear?) there. local ustring=unicode.empty(5) ;;ustring not unusable? Didn't think that out, just mindlessly copied ideas over from binary plugin. At the moment, that would be 5 WCHARs, all null (0x0000). So wcslen says it's empty (and at the moment I access validityy of a ustring[n] by making sure n fits within string length). but if ustring contained 5 WCHARs, all null, then you did ustring[3] = "E" you'd end up with \x0000\x0000E\x0000\x0000 (E being in its WCHAR representation) which is just the null string with an E buried after it. ???? > Also, it appeared I should be able to specify the type in the brackets when > setting only one character, but then I get position argument errors. With unicodePlugin0.73_090728.zip, you still getting problem? I'm not. > Maybe I'm doing something wrong? No, you're just using bits I forgot to test, as usual.
