Philippe>I know work on 3.0 is finished and milamber proposed to
be rm right ?

Sorry, I must have missed that.

Philippe>But if you want we can discuss later after 3.0

Anytime works for me provided it does not delay 3.0.

sebb>They don't care so much about memory footprint as they are not
benchmark utils.

I agree it is important to keep memory footprint sane, however I'm
sure it can easily be done negligible.

sebb> There's no particular reason to have it in the core tool.

You have no second chance to make the first impression.
Suppose a new user installs JMeter. How would (s)he know which plugins
are required?

Let me exaggerate to make the point more clear: "in order to install
JMeter you need to download ant, then perform svn checkout ..., then
ant download_jars, then ant install, then find JMeter Plugins in
Google, ..."

Not having out-of-the-box market makes first-time experience not that smooth.
Remember that typical JMeter users are not java programmers. They have
no-to-very little idea what the jar file is.
Of course, I've made that up, however I do believe that vision is right.


Having thought a bit more, I think jmeter-wrapper (named after "gradle
wrapper", though not using gradle in any way) script might be useful
as well (to have reproducible environment).
I mean the following:
1) User (or JMeter) creates a file with text description of exact
JMeter version and the versions of installed plugins. For instance:
jmeter-installation-versions.xml
2) If someone wants to reproduce exactly the same JMeter installation,
he picks that jmeter-installation-versions.xml
 and executes "./jmeter-wrapper".
3) The wrapper script downloads proper JMeter and plugins versions.

Vladimir

Reply via email to