>>>>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> On Thu, Apr 04, 2002 at 03:13:11PM +0200, Lars Gullik Bjønnes Andre> wrote: >> Note that the problems will lyxstring will only get bigger.... by >> clinging to it we introduce a lot of ineffencies (conversion >> between lyxstring-> char*->std::string and back... with all the extra >> allocations that result in), also because of these conversions >> strings cannot give any meaning to '\0' (and we Andre> I know. lyxstring was not my idea, though ;-) I began at some point to convert our stringstream implementation to be less GNU-ish (so that we can use it with lyxstring), but I gave up. It should not be too hard, though. And then we would always use this one with lyxstring. >> IMHO lystring is just looking less and less favorable. Andre> Which compiler that we support does not have a sufficient Andre> std::string? There are bugs in compaq cxx 6.2 with std::string (an extraneous assert), but I think it is fixed in 6.3 (cannot upgrade, though). Also lyxstring means much faster compiling and less disk footprint. I tought you were interested in these kind of reasons. Andre> /// Andre> Another thing: I just had a look at lyxlenght.[Ch]. Andre> Why the hack is it in such a state? I seem to remember making Andre> 'cleaning up patches' at least twice... I think what you sent has been applied. JMarc
