Scrap baloo?

[Dominik Haumann]

Hi,

On Wed, Sep 28, 2016 at 4:52 PM, Boudhayan Gupta <bgupta at kde.org> wrote:
> On 28 September 2016 at 02:36, Christoph Cullmann <cullmann at absint.com> wrote:
>> any update?
>
> Yep. In all the happennings of the week I just forgot to write this email.
>
> If Baloo is going to be an integral part of the Plasma experience, do

It already is. And that's exactly why we have this discussion. If you
look into baloo's bug tracker, almost *all* of the bugs are critical
crashes, and even of a kind that some people cannot use the desktop
anymore.

> we really want to depend on an external project where we don't have
> control (and indeed, sentiments may prevent unrestricted contributions
> based only on merit). This is the political reason why I don't want to
> depend against Tracker. The technical reason is that it's based on
> SQLite, which is incredibly slow compared to what we do now.
>
> At the same time, LMDB needs to be replaced, and fast.

Interestingly, this is not even necessarily the case: In NFS mounts,
sysadmins typically do not *want* an index, since it spams the network
load a lot. So not having it in an NFS environment is not the worst
option.

Further, 32 bit is going away. While the 1GB limit is not nice at all,
it's as simple as also not providing baloo there in case the db is
full. Given that most users are 64 bit meanwhile, this is not really
an issue. Still these are weak arguments for LMDB.

> I'm building a
> new KVDB as an university project (it should be able to do 256GB
> indexes on 32bit machines), and if that doesn't work out there's
> Sophia (http://sophia.systems/). I'll be evaluating both as a
> replacement to LMDB.

The reason why I bothered writing a mail is exactly this paragraph.

You also state that being able to do something crazy motivates you,
and that's the best motivation you could ever have :-) This is how KDE
came to life, and why it still lives today.

But trust me, writing a database on your own is neither easy nor
quickly done, and most certainly an issue maintainability wise.

Given we are looking for a solution now (in fact, better yesterday),
Christophs proposal is not that far away.

And: we can still switch away from tracker, think of it as a backend.
And in fact, I also believe that a shared index across desktops is a
nice-to-have. What we would need to agree on is probably that baloo is
a linux-only (?) framework then, but I guess that would possibly even
be fine.

Greetings
Dominik