jmeter-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philippe Mouawad <philippe.moua...@gmail.com>
Subject Re: svn commit: r1759062 - in /jmeter/trunk: bin/ src/components/org/apache/jmeter/validation/ src/core/org/apache/jmeter/gui/action/ xdocs/ xdocs/usermanual/
Date Tue, 06 Sep 2016 20:04:34 GMT
On Tuesday, September 6, 2016, Vladimir Sitnikov <
sitnikov.vladimir@gmail.com> wrote:

> Philippe>  - Caching with Guava, big discussion, patch refusal, 0 progress
>
> Just for the reference, here's recent discussion on Apache Calcite mailing
> list: "Dropping support for Guava versions earlier than 14"
> https://mail-archives.apache.org/mod_mbox/calcite-dev/
> 201609.mbox/%3CF40BEDBA-7180-4071-8DE9-03EF6439A3B6@apache.org%3E
>
> TL;DR: 18 messages in 3 days discussing how to approach "Guava problem"
>
> Re the CSS patch, I've not seen neither evidence of the problem (e.g.
> thread dumps or whatever) nor comparison (e.g. before vs after).
> So, I would play conservative here and decline adding questionable
> dependencies.

it's true about this particular case.
To be clear, I am not sure the bug is correctly labeled
https://bz.apache.org/bugzilla/show_bug.cgi?id=59934

I don't think there is a race condition in the sense that there would be a
threading bug.
It's just that in case of multiple threads loading the same css, many will
process it then cache while it would be more efficient to have only 1
process and cache and the others use the cache.

Did you try this case with 100 threads doing that ?

>
> If there was a case, I would be interested to fix that.
>
> PS. I prefer to discuss first, commit after even being a committer.
>
> Vladimir
>


-- 
Cordialement.
Philippe Mouawad.

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message