jmeter-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philippe Mouawad <p.moua...@ubik-ingenierie.com>
Subject Re: svn commit: r1840406 - in /jmeter/trunk: bin/jmeter.properties src/core/org/apache/jmeter/JMeter.java xdocs/changes.xml xdocs/usermanual/properties_reference.xml
Date Wed, 12 Sep 2018 19:27:42 GMT
I would stick with last Felix patch to respect the KISS principle.

1 property , and the language detection based on extension looks fairly
simple for me.

While the other one is not , it will require another bunch of documentation
and make code more complex only to be able to handle initialisation.

Why would one want to use 2 languages for initialising ? Let him choose 1
language.

Regards

On Mon, Sep 10, 2018 at 1:53 AM sebb <sebbaz@gmail.com> wrote:

> On 9 September 2018 at 16:16, Felix Schumacher
> <felix.schumacher@internetallee.de> wrote:
> >
> >
> > Am 09.09.2018 um 16:31 schrieb sebb:
> >>
> >> On 9 September 2018 at 14:55, Felix Schumacher
> >> <felix.schumacher@internetallee.de> wrote:
> >>>
> >>> Am Sonntag, den 09.09.2018, 14:41 +0100 schrieb sebb:
> >>>>
> >>>> On 9 September 2018 at 14:19, Felix Schumacher
> >>>> <felix.schumacher@internetallee.de> wrote:
> >>>>>
> >>>>> While I like groovy, it might be that other users have other JSR223
> >>>>> languages as their favourites.
> >>>>
> >>>> +1
> >>>>
> >>>>> Should we make the init script mechanism a little bit more flexible
> >>>>> by using
> >>>>> the files ending to determine the language/engine? That way a user
> >>>>> could
> >>>>> setup a jsr223.init.file=my_suberb_init.js and JMeter would try
to
> >>>>> get a JSR
> >>>>> 223 engine for "js", or jsr223.init.file=some_python.py to run it
> >>>>> with an
> >>>>> engine for "py".
> >>>>>
> >>>>> What do you think?
> >>>>
> >>>> To make it truly generic, I think the startup code needs to be given
> >>>> a
> >>>> list of languages to initialise.
> >>>> After all, someone might want to use two different JSR223 languages.
> >>>>
> >>>> One way to do this would be to have the following properties:
> >>>>
> >>>> jsr223.init.languages=groovy,js
> >>>> groovy.init.file=def
> >>>> js.init.file=abc
> >>>
> >>> Yes, that might be a good idea.
> >>>
> >>> But I think I might like it better if jsr223.init would be kept in
> >>> front of those variable names.
> >>>
> >>> What do you think of
> >>>
> >>> jsr223.init.languages=groovy,js
> >>> jsr223.init.file.groovy=def
> >>> jsr223.init.file.js=abc
> >>
> >> This will change the behaviour for beanshell unless that is handled
> >> separately.
> >
> > The current behaviour is that beanshell is handled differently.
> >
> >>
> >> Will also need to be aware that BeanShell init could appear in two
> >> places with different property names and files.
> >
> > My English parser seems to be broken, can you elaborate a bit, what you
> mean
> > with this?
>
> At present there is the property beanshell.init.file
>
> AIUI you propose to have the additional property:
>
> jsr223.init.file.beanshell=abc
>
> What happens if both are defined but point to different files?
>
> Whereas with my original naming convention there is only one property name.
>
> I'm not saying my suggestion is perfect, but I think it has fewer issues.
>
> > Felix
> >
> >
> >>
> >>> I put the "file" after "init" to enable a potential scripting engine
> >>> named "languages" ;)
> >>>
> >>>> If the code ignored missing *.init.file properties, then
> >>>> jsr223.init.languages could default to beanshell in order to keep the
> >>>> existing behaviour.
> >>>
> >>> Well, yes the names would be more appropriate, when .init.file is the
> >>> postfix. Hm.
> >>>
> >>> Felix
> >
> >
>


-- 
Cordialement.
Philippe Mouawad.
Ubik-Ingénierie

UBIK LOAD PACK Web Site <http://www.ubikloadpack.com/>

UBIK LOAD PACK on TWITTER <https://twitter.com/ubikloadpack>

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