Re: [Dltk-dev] Re: HierarchyResolver performance improvement

[Alex Panchenko <alex@xxxxxxxxx>]

Hi Michael,

1) acceptType() originally had empty implementation, now you changed it to be abstract. I suggest to introduce new class with empty acceptType implementation, since there are several places where search request is issued just to warm up the search engine.

2) There are differences in type hierarchy mentioned below with and without this patch, so we should find out why, and fix the appropriate part of the code.

Regards,
Alex

----- Original Message -----
From: "Michael Spector" <spektom@xxxxxxxxx>
To: "DLTK Developer Discussions" <dltk-dev@xxxxxxxxxxx>
Sent: Saturday, April 18, 2009 5:45:55 PM GMT +06:00 Almaty, Novosibirsk
Subject: Re: [Dltk-dev] Re: HierarchyResolver performance improvement

Hi Alex,

On Fri, Apr 17, 2009 at 12:18 PM, Alex Panchenko <alex@xxxxxxxxx> wrote:
> Hi Michael,
>
> 1) I think it make sense to introduce
> public final class NopTypeNameRequestor extends TypeNameRequestor,
> since it is used in several places to warm up the search engine.

I'm not sure I understand what's the purpose of NopTypeNameRequestor.
Can you explain?

>
> 2) Type hierarchy does not look correct for Ruby (probably it was partially affected by https://bugs.eclipse.org/bugs/show_bug.cgi?id=260215 related changes, but the proposed changes add additional incorrectness too), so more time is required to review/improve these changes.
>
> a.rb:
> class AAA
> end
>
> b.rb:
> class BBB < AAA
> end
>
> and call hierarchy for BBB.

You mean: multiple packages/modules are shown in TypeHierarchy view?
This looks like a UI problem. Hierarchy produced by hierarchy resolver
looks OK...

>
> I hope to find time to look it thoroughly during weekend or on Monday.
>
> Thank you,
> Alex
>
> ----- Original Message -----
> From: "Michael Spector" <spektom@xxxxxxxxx>
> To: "DLTK Developer Discussions" <dltk-dev@xxxxxxxxxxx>
> Sent: Thursday, April 16, 2009 10:08:15 PM GMT +06:00 Almaty, Novosibirsk
> Subject: [Dltk-dev] Re: HierarchyResolver performance improvement
>
> Sorry, here's the correct link:
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=272513
>
> On Thu, Apr 16, 2009 at 6:07 PM, Michael Spector <spektom@xxxxxxxxx> wrote:
>> Hi All,
>>
>> Please see this bug: https://bugs.eclipse.org/bugs/show_bug.cgi?id=272423
>> I've attached a patch that dramatically improves the hierarchy
>> resolver performance. If none has any objections - I'll commit it.
>>
>> PS: there's a small API change, I've changed all parts that depend on
>> TypeNameRequestor.
>>
>> Thanks in advance.
>>



-- 
Best regards,
Michael
_______________________________________________
dltk-dev mailing list
dltk-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/dltk-dev