Jean-Marc Lasgouttes wrote:
Mael Hilléreau <[EMAIL PROTECTED]> writes:
Some users use comments, notes to put quick reminders which shouldn't
be spellchecked at all. Other users will e.g. use notes as drafts and
could like them to be spellchecked. Only a per-inset or per-layout
approach can address these needs.
I do not think we should try to cater for each and every possible
pattern of use of lyx. I think we are in over engineering mode. We
should implement a completely manual character-level no-spellcheck and
be done with it.
JMarc
Yes, I agree. I've actually started to implement this, but ran into the
following question:
On the one hand, I want to just add this attribute to the Font (together
with language, for example, as a bool property) --- mainly in order to
utilize the Font-span mechanism.
OTOH, I'm not sure that it really belongs within the font. More
specifically, if it's part of the font, then should font comparison (==
op) be affected by this? Should this attribute be copied when the font
is copied? etc... Note that if this were implemented as a "none"
language, then the answer to both of the above questions would have been
"yes", so perhaps it's not so bad (since that ideas seemed to be
generally acceptable, except for the problem of loss of language
information).
Finally, another possibility (still within the font, though) is to add
it to the FontBits struct -- but I wasn't sure what the TOGGLE, IGNORE
and INHERIT flags are, or if they're even relevant here. Could anyone
pleas explain these?
Thanks!
Dov
