|
|
|
|
|
|
|
|
Re: color scheme for java editor [message #60113 is a reply to message #59332] |
Wed, 18 June 2003 05:44  |
Eclipse User |
|
|
|
Please see my comments below.
"Vladimir Blagojevic" <vladimir@cs.yorku.ca> wrote in message
news:bcne5f$ljo$1@rogue.oti.com...
> > yes, they could
> > the problem is that parsing a java file takes, let's say 300ms
> > that might be too slow for smooth syntax highlighting
> > what you need is incremental AST, and that we don't have
> > see https://bugs.eclipse.org/bugs/show_bug.cgi?id=36890
>
> Thanks Adam. I wish the reporter (Jerome Lanneluc) was a little bit more
> explicit. It is a nice practice to describe the problems thoroughly :)
This bug was not intended for public consumption, but more as a reminder for
a possible work item.
>
> I am guessing that Jerome wants to say that rebuilding the whole class
tree
> after one of its nodes changes is deemed too expensive and inefficient?
Right.
> Sure, that is going to help somewhat I guess. But you still have to
> construct the syntax tree for the whole java file before syntax colouring
> it? How is incremental AST going to hep in that case?
If incremental AST is associated with a callback mechanism, then the syntax
coloror could listen to these events and redraw only the parts that have
changed.
>
> Or does Jerome want to have an AST consisting of lazily created nodes, so
> that only source visible part of the java file has the actual AST node
> created and the rest (invisible part) consist of some sort of lightweight
> node proxies? I think that this is what he wants when he talks about
"create
> nodes on request".
Thats' correct.
>
> Best regards.
>
>
|
|
|
Powered by
FUDForum. Page generated in 0.07097 seconds