Hi Rob,
I thought about it longer and maybe we have to reverse our strategy.
Instead wanting to build a class explorer out of Scintilla, we should
do the opposite. Look what Scintilla does and shape that in an
interesting form. Basically what Scintilla does and which is
interesting or an editor is lexing and folding.
* folding
Folding is about hierarchy. So instead of a class explorer we could
rather think of a 'fold explorer' in which we group all folding nodes
in a tree. This offers another hierarchically view on the code, which
is not necessarily less interesting than a class explorer. It is maybe
something new, but can be very handy when you have to explore a
programming language or xml. Scintilla gives it all for free. The
problem of real class explorers is that they are language specific and
that Scintilla does not provide enough information without resorting
to writing custom scripts for every language.
What could be an interesting possiblility is that if this fold
explorer (as it will have start and end positions) supports drag and
drop, so you can reorganize your code with a click.
* lexing
I see lexing more as a way to identify. We could build a list of all
words which are styled from which we could generate all kinds of
lists, which can be represented in list controls with columns. There
could be options to turn certain styles (classes, methods, keywords,
...) on or off.
Folding and lexing can also be combined to provide the folding nodes
with style tags, which makes all kinds of visibility filters possible
for the 'fold explorer'.
For retrieving the lexing/styling information you have already a
prototype. A prototype to find folding nodes with their hierarchies
should also be doable.
With these two prototypes we can develop two minors: 'fold explorer'
and 'style index'.
What do you think?
Stani
On 3/1/07, SPE Stani's Python Editor <[EMAIL PROTECTED]> wrote:
> Hi Rob,
>
> Good job!
>
> Concerning the hierarchy in SPE that is most simple part of the
> answer. As said before I use my own custom classes:
> realtime.TreeCtrl
> realtime.ListCtrl
>
> Their API is (nearly) identical to the wx.TreeCtrl and wx.ListCtrl.
> There is a small, subtle difference in resetting them, but don't worry
> about that. So I would suggest you implement it with normal a
> wx.TreeCtrl (for hierarchy) or wx.ListCtrl (for sorted list). If you
> manage to do that I'll help you porting them to my realtime classes if
> you want to use them.
>
> The lexer gives already a lot of information, but as you noticed it is
> not enough for a tree. However maybe you should start with an index
> list which can be sorted by its columns (by name, style or line
> number). I think that is step 1.
>
> Step 2 would be trying to find out the hierarchy. This is more
> difficult, not so much for python, but to make it work in a generic
> way for all the lexers is more difficult. For that probably we need to
> use the folding information as well.
>
> These are some issues I immediately notice, although I am not expert:
> - the style bits are not the same across languages as in some
> languages classes are not even highlighted (C++) For example:
>
> * python
> # Class name definition
> style.python.8=fore:#0000FF,bold
> # Function or method name definition
> style.python.9=fore:#007F7F,bold
>
> * C++
> # UUIDs (only in IDL)
> style.cpp.8=fore:#804080
> # Preprocessor
> style.cpp.9=$(colour.preproc)
>
> * Pascal
> # Symbols
> style.pascal.8=fore:#007F7F
> # Preprocessor
> style.pascal.9=fore:#7F7F00
>
> But languages as Python and Ruby do share the same style bits. So
> maybe it boils down grouping the supported languages by Scintilla in
> groups, which still might be less work than writing it yourself for
> every language. (The grouping could be done automatically by mapping
> the style bits with their labels in dictionaries and scan for
> similarities.) But as classes and functions in C++ are not
> highlighted, how can you extract them? Maybe you or someone else
> (anyone?) on this list has better ideas. Otherwise it could be good to
> ask Neil if this is utopia or realistic.
>
> Even if this approach doesn't work it still would be good to develop a
> generic hierarchic parser api, for which maybe with regular
> expressions we can write plugins.
>
> Stani
>
> On 2/25/07, Rob McMullen <[EMAIL PROTECTED]> wrote:
> >
> > Stani,
> >
> > I'm looking through your explorer implementation and have incorporated
> > it as a minor mode in peppy. It was pretty straightforward to
> > incorporate, although I haven't grokked the internals yet. Still
> > reading the code. :)
> >
> > I threw together this code that looks at the stc and grabs anything
> > that is of the requested styles -- it outputs a list of tuples: (start
> > pos, end pos, style type, and text). This is the totally naive
> > approach of just marching through the text and gathering styles that
> > are marked as the same.
> >
> > In your Panel, for example, I added this method:
> >
> > def updateExploreGeneric(self):
> > classes = {wx.stc.STC_P_CLASSNAME: 'Class definition',
> > wx.stc.STC_P_DEFNAME: 'Function or method',
> > }
> > getLexedItems(self.source, classes)
> >
> > that calls this function:
> >
> > def getLexedItems(stc, classes):
> > length = stc.GetTextLength()*2
> > text = stc.GetStyledText(0, length)
> > bits = (2**stc.GetStyleBits()) -1
> > print "seaching for %s" % classes
> >
> > i=1 # styling bytes are the odd bytes
> > parsed = []
> > while i<length:
> > # get the style, stripping off the the indicator bits
> > style = ord(text[i]) & bits
> > if style in classes:
> > # it's a style we're interested in, so gather
> > # characters until the style changes.
> > found = i-1
> > i+=2
> > while i < length:
> > s = ord(text[i]) & bits
> > if style != s:
> > parsed.append((found, i-1, style, text[found:i-1:2]))
> > break
> > i+=2
> > else:
> > i+=2
> > print parsed
> >
> >
> > So... It's a start. I've got to figure out your hierarchies and I'll
> > see if I can merge a generic implementation into your Panel.
> >
> > Rob
> >
> > > >
> >
>
>
> --
> http://pythonide.stani.be
> http://pythonide.stani.be/screenshots
> http://pythonide.stani.be/manual/html/manual.html
>
--
http://pythonide.stani.be
http://pythonide.stani.be/screenshots
http://pythonide.stani.be/manual/html/manual.html
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"pyxides" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.nl/group/pyxides?hl=en
-~----------~----~----~----~------~----~------~--~---
