[ jEdit-devel ] Combining folding modes

[Shlomy R. <sre...@gm...>]

Hi,

I created a small plugin named FoldTools, which is supposed to provide
several features related to folding modes.
One of the features is something I found very useful - a "Composite" fold
handler which combines several folding modes. For example, a "comment"
folding mode, which folds all multi-line comments, with a "custom" folding
mode that folds all code blocks (lines between '{' and '}' for example).
Combining both, I can fold comments and blocks together in a buffer. This
 works well if the combined folding modes do not conflict.

However, there's a small detail that is missing to make the combined mode
work well. There are two types of folding modes:
1. Modes where the fold level of line 'n' depends only on the content of
line 'n'. (e.g. indent folding mode)
2. Modes where the fold level of line 'n' depends on the content of line 'n'
and on the fold level of line 'n-1'.

In order to combine several folding modes, the Composite fold handler must
know which type is each of the combined folding modes. Due to the fact that
each of the existing folding modes is meant to be used "standalone", the
Composite fold handler should only use the difference between the value that
each mode provides for the current line and some value that  is provided for
the previous line. For modes of type (1) above, this should be value
returned by the mode for the previous line. For (2), this should be the
value that was returned by the Composite fold handler for the previous line.

Currently I keep a fixed map in the plugin, for all known folding modes,
that indicates which type each mode is. But to make this more generic, it
would be useful for a FoldHandler to also return this information. What do
you think about adding another method to FoldHandler - a boolean
"isContextSensitive()" or "dependsOnPreviousLine()"?
Then I can remove this fixed map and can make sure it works correctly with
future modes.

Thanks,
Shlomy