[Kde-pim] Marketing blocker collection: Searching not Reliable

[Vishesh Handa]

On Tue, Mar 12, 2013 at 1:25 AM, Lindsay Mathieson <
lindsay.mathieson at gmail.com> wrote:

> Sorry I can't be more specific than this, I've been trying to narrow it
> down
> for weeks but there's just so many issues.
>
> Test Bed is a Gmail account with around 20,000 msgs. I've removed the
> account
> and re-added with disconnected mode enabled and a clean nepomuk index.
>
> I've forced a nepomuk reindex from scratch many times. I've built clean
> VM's
> and let them sync and index over the course of days. All of them fail
> searching badly.
>
> - Inaccurate. Searches are just not finding messages that match simple
> criteria. It seems to be worse in some folders than others. I have folders
> where a subject search on "" returns only half the messages in the folders,
> others were its only missing a few.
>

Could you please list examples of where the search was inaccurate? If I can
reproduce it then I can probably fix it.


> - Inconsistent. Searching on root folder + subfolders returns X msgs.
> Searching on All returns Y msgs. Y is always larger than X and seems to
> consist of folders that weren't found in X.
>

Hmm. I'll try it out.


>
> - Doesn't update. Msgs sometimes take hours to get into the index.
> Sometimes
> they never do. I see this often with "Msg Status = Unread".
>

Christian has pushed some improvements to the feeder already. And I'm
planning on integrating the nepomukcontroller to show some kind of progress
window or something. You'll basically be able to see how many emails
haven't been indexed.


> - Saved Searchs update erratically, if at all.My std Saved Search is "Msg
> Status = Unread". This is always massively out of date within a day. The
> same
> search run manually shows a completely different result.
>

This won't happen any more. They won't update at all :D

Sorry, it was too cpu consuming to keep it enabled. It should be disabled
with 4.10.2. I'll try and fix this properly with 4.11. Though I'm kinda
surprised that the results were not up to date. Could you give me an
example query?


> Basically searching is completely broken. My guess is that its the
> indexing.
>

Actually, the indexing is pretty decent right now. It might just be the
searching part. We haven't really tested it much.


>
> It needs to be close to if not at 100% accurate to be useful, currently it
> not
> even at 50%.
>
> On the plus side :) where its working its *very* very* fast. Full text
> search
> on a 12,000+ folder is an order of magnitude faster than Thunderbird. But
> were
> only talking seconds, so its not a difference the average user will care
> about.
>
> It does do the initial sync of msgs much faster than Thunderbird.
>
>
>
> --
> Lindsay Mathieson
> _______________________________________________
> KDE PIM mailing list kde-pim at kde.org
> https://mail.kde.org/mailman/listinfo/kde-pim
> KDE PIM home page at http://pim.kde.org/
>



-- 
Vishesh Handa
_______________________________________________
KDE PIM mailing list kde-pim at kde.org
https://mail.kde.org/mailman/listinfo/kde-pim
KDE PIM home page at http://pim.kde.org/