Issue with KDevelop: Is it using different parsers in parallell for different puropuses?

mwoehlke mwoehlke at
Fri Aug 25 15:39:23 UTC 2006

Kuba Ober wrote:
>>> KDevelop 4 won't have a new editor.  It is using the KTextEditor
>>> interfaces. Those interfaces allow KDevelop 4 to choose to do its own
>>> highlighting.  The only way this could be re-used for KWrite is for the
>>> default KTextEditor implementation -- katepart -- to use kdevelop-pg
>>> parsers for its highlighting.  Who knows, perhaps the Kate devels will
>>> eventually do this, but the place to talk about that is on kwrite-devel
>>> as you yourself noted.
>> True. Well... since obviously the decision is made, I guess I'll wait
>> and see. I still think it would be good if KATE's highlighting was more
>> language aware. It will be unfortunate if KATE winds up implementing a
>> different language-aware parsing engine that competes with KDevelop's.
> Kate already has it. It's implemented and I don't see it being reworked any 
> time soon, unless a volunteer steps up to the plate. Calling it a parsing 
> engine is buzzowordiness to me. It's a simple state machine coupled to a 
> pattern matcher. Simplicity is the key word here.
> It works pretty well considered how simple it is. It *is* langauge aware.
> [snip]

"Grammar" aware; KATE doesn't know about grammars. The point would be to 
have MUCH better error detection, primarily, but I think it would be 
useful in other ways as well. But as you say, SHTDI and it isn't likely 
to happen any time soon. However, since it no longer sounds like 
KDevelop is implementing a full-fledged grammar-based highlighter 
(instead they are doing something that works /with/ KATE, not /instead/ 
of KATE), there is no urgency.

> Have you actually tried what you're talking about?

I know it works *now*... it was sounding like it was going to *stop* 
working, which is what I was worried about. :-)

> The only thing, I gather, that KDevelop may potentially do is override the 
> highlighting for files it knows how to highlight better than what Kate does. 
> That more powerful highlighting functionality could be potentially put into a 
> separate framework and made useable by Kate (and other applications), but 
> again -- someone has to design it and code it first. I suggest you give it a 
> try if you place such a high value on that feature (as you claim).

If you read the further posts, it sounds like this is NOT what KDevelop 
is doing. And what I place high value on is not reinventing the wheel 
somewhere that it can only benefit a subset of users. But this is not 
what is happening. Please go back and read the rest of the thread.

'$ time make world' -> real 5d:14h:37m:5.291s  user 0m:0.000s  sys 

More information about the KDevelop-devel mailing list