John Ehresman:
Did run-length encoding save space in many cases? I realize that in some situations it will probably take more space.
One byte run lengths with one byte styles is roughly equivalent to current interleaved styling for common programming languages. You can tweak the style definitions to, for example, use a single style for white space and identifiers and this decreases the number of runs so its a win although there is then the extra time taken for processing runs so its a tradeoff. The current implementation is actually positions and values rather than lengths with some of the positions being relative to the most recent modification so they don't churn too much. The motivation for adding this feature was for features like showing matches, URL highlighting, the selection and similar.
I'm also interested in stacked styling -- one use case is allowing the view to style individual characters (to show search matches, for example).
Yes, that is the sort of thing where you place an extra low-density styling buffer (like the above) over the main styling buffer. There are complexities in determining how the styling buffers combine into a single image: some (like underlines or boxes) may be drawn after the main text but others (such as making the background redder or drawing the text as bold) need to look at all of the buffers and have some idea of priority. Neil _______________________________________________ Scintilla-interest mailing list [email protected] http://mailman.lyra.org/mailman/listinfo/scintilla-interest
