[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kwrite-devel
Subject:    Re: [Bug 154810] word completion: detect filetype and populate word
From:       Anders Lund <anders () alweb ! dk>
Date:       2007-12-30 23:02:14
Message-ID: 200712310002.14545.anders () alweb ! dk
[Download RAW message or body]

On Sunday 30 December 2007, Thomas Friedrichsmeier wrote:
> On Sunday 30 December 2007, Anders Lund wrote:
> > On Sunday 30 December 2007, Thomas Friedrichsmeier wrote:
> > > Therefore, the first wish would be the ability to force a *specific*
> > > version of a highlighting defintion to be used, e.g. by filename.
> >
> > It is possible to select the highlight using the configuration dialog,
> > either by setting priority, by defining it in a filetype of by using a
> > document variable, possibly in a .kateconfig file. I think that covers
> > that.
>
> I'm not entirely sure, we're talking about the same thing, here. My concern
> is more about IDEs that want to do *more* than what can easily be fit into
> a highlighting definition (but at the same time only deal with a limited
> number of different file types). For instance kdevelop might want to add
> some specialities to C/C++ files, quanta to php / html files, RKWard to R
> files, etc. In that situation it might be useful for those applications to
> control the highlighting definition in use more closely (i.e. independently
> of user settings or kate filetypes). For those cases kdevelop/quanta/rkward
> might want to make sure that a very specific version of the highlighting
> defintion is used, independently from any version(s) of the highligthing
> defintion the user might have installed elsewhere.
>
> (But then again, maybe you are right and this could be solved by shipping
> an own highlighting definition which simply sets a high priority?)
>
> > > Further wishes:
> > > - get and set <list/>s of keywords
> > > e.g. void setKeywordList(const QString &listname, const QStringList
> > > &list);
> >
> > I don't know why we would want to edit the highlight files this way.
> > Afterall the user is blind in this case, as she can't possibly know for
> > sure how the list is used.
>
> Yes, but see above, again. kdevelop might want to provide code completion
> for preprocessor directives, quanta might want to make sure additional PHP
> functions which are defined in the project get highlighted, RKWard might
> want to add highlighting for functions / symbol names loaded from arbitrary
> packages. This is possible by implementing arbitrary highlighting rules,
> but it would simply be very convenient to be able to pass on the job to the
> kate highlighting engine, which already deals with those same aspects, and
> just refine/reuse the data in some cases.

IDEs does not want to use our simplistic tools, they are way more advanced and 
spphisticated ;)

> > > - on a more basic level, the ability to find out at least, whether a
> > > specified cursor is inside a comment region, or perhaps in a string
> > > literal.
> >
> > This would be covered  by the above if i am not wrong? It can currently
> > be done by checking the attribute.
>
> Yes, this would be covered by the above (now snipped, though it actually
> would be the most important feature). This is just a more limited
> suggestion, in case the above is too hard to accomplish.
>
> How would I get at this attribute, right now?

There was a method in the now gone highlight interface, I believe it's now in 
smartrange. But again, things might have changed that I am not aware of.

> > > - ability to get and set the regexp defining a "word". See
> > > http://lists.kde.org/?l=kwrite-devel&m=119400167931455&w=2 .
> >
> > First, there is no regexp, rather a list of characters defined as word
> > separators. And it's global to the highlight i believe.
>
> I guess you may be right. Inside
> KateCompletionWidget::cursorPositionChanged(), a regexp is being used,
> though (and perhaps that is the point where it should be addressed,
> instead).

That is wrong, imo. It should use the function provided by katedocument. I 
don't know the reasoning behind it though.


-- 
Anders

www: http://www.alweb.dk
jabber: anderslund@jabber.dk
_______________________________________________
KWrite-Devel mailing list
KWrite-Devel@kde.org
https://mail.kde.org/mailman/listinfo/kwrite-devel
[prev in list] [next in list] [prev in thread] [next in thread]