Jeremias Maerki wrote:
I'd prefer to make the area tree more detailed and the renderers
Maybe we could use a trick to minimize the memory increase by creating
an additional "word" area as a child to the text area so we don't have
to replicate the traits for each word. WDYT?
So, your idea is to move the actual string of text from TextArea to a new
class WordArea; in the pdf renderer, renderText would just create the
parameters and the TJ operator, while the rendering of a WordArea would
produce a string (...) or <...> with its offset, right?
Tomorrow I'll try and implement this.