Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [jdt-core-dev] [jdt-dev] Discussion about the proposed JDT index

Sure thing, Dani. Let me know where and when.

We may want to consider running a hangout in addition to Asterisk (which I assume still just does audio) since that would let us screen-share documents, but I'll go along with what you decide if you want to take over organizing this.

On Fri, Nov 13, 2015 at 7:52 AM Daniel Megert <daniel_megert@xxxxxxxxxx> wrote:
> I'm going to delay this until Manoj can attend,

Thanks Stefan. I was about to ask for that here. It makes sense to wait for Manoj since he is currently the owner of search and the index(er). Given the amount of interested parties I suggest to send a Doodle.

I'm all in for making the index creation and usage faster, and any help is appreciated. Having a kickoff call is a good start for this, and I will join (GMT+1). I suggest that we use Eclipse infrastructure (Asterisk) to have the call. I can provide the details if that's fine for you.

Dani



From:        Stefan Xenos <sxenos@xxxxxxxxxx>
To:        "Eclipse JDT general developers list." <jdt-dev@xxxxxxxxxxx>, Sergey Prigogin <sprigogin@xxxxxxxxxx>, Lars Vogel <lars.vogel@xxxxxxxxxxx>, jdt-core-dev@xxxxxxxxxxx, Doug Schaefer <dschaefer@xxxxxxx>
Date:        13.11.2015 16:35
Subject:        Re: [jdt-dev] Discussion about the proposed JDT index
Sent by:        jdt-dev-bounces@xxxxxxxxxxx




I'm going to delay this until Manoj can attend, but please continue sending me your time zone and gmail address if you wish to attend and I'll follow up.

On Wed, Nov 11, 2015 at 7:25 AM Stefan Xenos <sxenos@xxxxxxxxxx> wrote:
Hello!

As some of you know, I've been working on a proposal for a faster JDT index. I'd like to organize a google hangout to answer questions about the proposal, understand the needs of the JDT core committers, and decide on the next steps to take for development.

I'll make myself free any time over the next week. Anyone is invited. I'd very much like it if we could get myself (GMT-5), Sergey Prigogin (GMT-8), Dani Mergert (GMT+1?), Stephan Hermann, and Sebastian Zarnekow in attendance. Due to time zone differences, it's going to be at an awkward time for someone, but I'll work around the schedules of the JDT committers.

If you'd like to attend, please let me know. Suggest a time at which you'd be available over the next week, your hangouts username (aka gmail address) and whether you have any items you'd like to add to the agenda.

Of interest to me are:

- Does anyone have questions or concerns about the design and implementation so far?

- Indexing time. A number of folks expressed concern about indexing time due to the lack of prebuilt indices. I suggested that the use of a nonblocking indexer permitting parallel reads makes indexing time much less important than it was in the past. Let's work this out.

- How can we collaborate on the index? We've been using a branch in a google-internal git repository. I've dumped a snapshot of the branch into a gerrit patch, but this isn't very transparent and doesn't facilitate collaboration. Let's come up with something better.

- Stephan drew attention to the PDOM prefix. I'm also not a fan of of the all-caps acronym in the class names. Can we come up with something better? Are there any other names, packages, or stylistic issues we're concerned about?

- What is the most important item for me to focus on next? Possibilities:
  - Polish the current limited use-case (type hierarchy view searches) so that we can get it into master and start using it for something real.
  - Update the rest of the type hierarchy view to use the new index so we can see how fast it runs when backed entirely by the new index.
  - Focus on the indexer and filling in the remaining data so we can get better numbers about indexing time and index size.
  - Adopt the index in the call hierarchy view so we have more diverse use-cases.
  - Benchmarks and performance tests

- What are the acceptance criteria that must be met before this can go into JDT core?

- Which JDT core committer(s) will be willing to help us with code reviews?

- Is anyone interested in helping with the implementation or adoption? If so, let's gather up the names. I'll follow up with them offline.

Bug: https://bugs.eclipse.org/bugs/show_bug.cgi?id=481796
Design doc: https://drive.google.com/open?id=1w3-ufZyISbqH8jxYv689Exjm0haAGufdcSvEAgl2HQ4
Implemenation: https://git.eclipse.org/r/60004

  - Stefan_______________________________________________
jdt-dev mailing list
jdt-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev

_______________________________________________
jdt-core-dev mailing list
jdt-core-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-core-dev

Back to the top