Re: [PATCH 1/4] utils: Change critical section terminology to worker threads.

[Christopher Baines <mail@cbaines.net>]

[-- Attachment #1: Type: text/plain, Size: 1443 bytes --]


Ludovic Courtès <ludo@gnu.org> writes:

>> This doesn't mean all queries have to be executed one at a time though,
>> providing the queries are executed outside the threads used by fibers, and a
>> single connection isn't used in multiple threads.
>
> Serialization of database accesses comes from:
>
>   commit 4612a3a70f1e70afa4e0ce00e8cb1a7848dddf58
>   Author: Clément Lassieur <clement@lassieur.org>
>   Date:   Sun Aug 5 13:14:44 2018 +0200
>
>       database: Serialize all database accesses in a thread.
>
>       Fixes <https://bugs.gnu.org/32234>.
>
> Apparently at the time we weren’t quite sure how SQLite would perform if
> we accessed it from several threads, but you seem to suggest that it
> works fine, right?

Yeah, the documentation suggests it can work [1]. We're even opening the
database in multi-threaded mode already (SQLITE_OPEN_NOMUTEX).

1: https://www.sqlite.org/threadsafe.html

> Did you try it on a big database like that of berlin?

I have an old copy of the berlin database, and it seems to work fine
with that. At least the web interface that is, but if that works,
everything else should work too.

>> These changes start to move in this direction, first by just changing the
>> terminology.
>
> That’s also welcome!

I've gone ahead an pushed these patches now, as it's something I'd like
to deploy prior to the Guix days.

Thanks for taking a look,

Chris

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 962 bytes --]