rgheck wrote:
Jean-Marc Lasgouttes wrote:
rgheck <rgh...@bobjweil.com> writes:
So we have two different Inset*::getCaptionText methods with different
signatures? This looks ugly, I'd rather standardize on one.
I've changed the name to "getCaptionAsPlaintext()", which is what it
does.
And is it warranted to have one with strings and one with streams?
Maybe I'm missing something, but which "one with strings" do you have
in mind?
Oh, sorry, I see what you mean. We have:
docstring InsetText::getCaptionText(OutputParams const &) const
and
int InsetCaption::getCaptionAsPlaintext(odocstream & os, OutputParams
const &) const
Having looked at how these are used, of course you could change the
signature of either, but they are used to do very different things, and
they really make most sense as they are. I think. Changing the former
would just force the odocstringstream::str() call to the client;
changing the latter would introduce ugly indirection, since this is
called from InsetCaption::plaintext() (where we have the odocstream) and
wants to call InsetText::plaintext, which needs an odocstream.
rh
