I guess you're right. My goal was to see how rodauth-oauth acts under 
concurrency, but I might have gone beyond what I wanted to test.

I ended up disabling parallel testing for jdbc sqlite due to what I found 
here: 
https://stackoverflow.com/questions/10707434/sqlite-in-a-multithreaded-java-application
 
.

Thanks again for the support,
Tiago

A domingo, 30 de agosto de 2020 à(s) 18:58:36 UTC+1, Jeremy Evans escreveu:

> On Sunday, August 30, 2020 at 8:45:06 AM UTC-7, Tiago Cardoso wrote:
>>
>> > Are you using a connection string of "jdbc:sqlite::memory"?
>>
>> In the link I sent you, I used "jdbc:sqlite:/$(pwd)/test.db. The error is 
>> related with some file lock, which probably means that access to the file 
>> db in sqlite is assumed not to be concurrent(?) and running tests in 
>> parallel can't possibly go?
>>
>> > You could try limiting the ActiveRecord and Sequel connection pool 
>> sizes to 1
>>
>> I'll try that, thanks for the suggestion.
>>
>> > Unless it is critical, I would recommend not testing in parallel mode 
>> when using SQLite,...
>>
>> That's definitely on the table. However, concurrent tests don't break 
>> with CRuby, hence why I thought something was wronng with the jruby build. 
>> Maybe the c-extension is somehow limitng the pool?
>>
>
> Maybe.  It seems unlikely anyone is using JRuby/SQLite/rodauth-oath in 
> production, which is why just making the tests single threaded seems 
> easiest. If you are on JRuby/JDBC, a database written in Java such as 
> H2/HSQLDB/Derby is probably a better idea than SQLite if you did want a 
> file-based database.
>
> Thanks,
> Jeremy
>

-- 
You received this message because you are subscribed to the Google Groups 
"sequel-talk" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sequel-talk/62694125-a7fb-4c7a-a251-773d3e4e1321n%40googlegroups.com.

Reply via email to