Andre Poenitz wrote: > On Fri, Jul 18, 2003 at 04:03:48PM +0100, John Levon wrote: >> On Fri, Jul 18, 2003 at 04:03:09PM +0200, Andre Poenitz wrote: >> >> > > 'switching to unicode' means reading in, writing out >> > >> > Rather some 8 bit encoding of it like UTF8(?).. >> >> Talking to Lars yes I think we want UTF8 as the file-format encoding.
This bit I understand. >> Storage at runtime is more controversial. Lars wants to use the >> platform's wstring etc. I think this will cause us a lot of problems - >> support varies wildly between different systems, and somebody is going >> to have to write the wstring <-> QString stuff. I take it that conversion to/from a QString would be trivial were we to use a 16bit type internally (given your assertion that QString is a UTF-16 data buffer) but that we would need something more complex if we stored the data using a 32bit type? >> I placed everything behind a to/fromqstr buffer zone precisely so this >> can be done if still needed. > > What about xforms? We'd need something like this there as well.. xforms uses XmbLookupString etc internally. I guess that this means LyX's internal store would be converted to an 8 bit encoding which xforms could then handle. -- Angus
