Angus Leeming <[EMAIL PROTECTED]> writes:
| Lars Gullik Bjønnes wrote:
>> | No it's not. If the data is passed as a string then we allow the
>> | outside world to receive the data, modify it and post it back with
>> | no addditional effort to us.
>>
>> this is the exception (external data), for all internal stuff
>> serialization _and_ the mailer stuff are really overkill.
>
| Shrug. There was a simple scheme that worked in 1.3.x where the
| frontends stored the inset. Which you didn't like. Now you have a
| safe scheme that you think is too complex. Yet I would argue it is
| clean and simple and extensible.
>
| Come up with a working alternative and we'll discuss that ;-)
Simple: forget about the externals for a second.
Pass the whole InsetParamsXXX in a boost::any to where it is used. The
sender and the receiver knows the type so now prob. Now all the mailer
code can go.
| Nonsense. We have three frontends now. Anything modified by the
| frontend should be regarded as suspect. As suspect as anything coming
| from the lyxserver.
I don't agree.
| Incidentally, I believe that boost::serialization does not support
| polymorphisism. Which leaves inset->read() dead in the water.
I am not sure that we would need it anyway...
--
Lgb
