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: New feature : Readable overview of Test Plan
Date Tue, 22 Jan 2019 20:58:44 GMT
Hi Felix,
I added patch so you can have a look to  it:
https://bz.apache.org/bugzilla/show_bug.cgi?id=63101

I think it is highly "improvable".
For now I put in Menu Item (alpha), this way it can easily be tested.


Regards

On Tue, Jan 22, 2019 at 9:03 PM Felix Schumacher <
felix.schumacher@internetallee.de> wrote:

>
> Am 22.01.19 um 20:36 schrieb Philippe Mouawad:
> > Hello,
> > Do you think this feature as of now (based on XSL) is worth being
> committed
> > to core as alpha?
>
> That depends on how alpha you think it is :)
>
> Maybe include it, but disable it by default with a clear warning, that
> it is there for testing the usability?
>
> >
> > In terms of implementation, it relies on Saxon BasicTransformer as it
> > supports xslt 2.0 functions.
>
> Are the used functions comfort functions or necessary ones? (Just out of
> curiosity).
>
> Felix
>
> >
> > Regards
> >
> > On Monday, January 21, 2019, Felix Schumacher <
> > felix.schumacher@internetallee.de> wrote:
> >
> >> Am 21.01.19 um 20:56 schrieb Philippe Mouawad:
> >>
> >>> On Monday, January 21, 2019, Felix Schumacher <
> >>> felix.schumacher@internetallee.de> wrote:
> >>>
> >>> Am 21.01.19 um 11:40 schrieb Philippe Mouawad:
> >>>> Hello,
> >>>>> This will better explain feature:
> >>>>> Menu:
> >>>>>
> >>>>>       - http://ubikloadpack.com/demo/schematic/MenuSchematicView.png
> >>>>>
> >>>>> Plan:
> >>>>>
> >>>>>       - http://ubikloadpack.com/demo/schematic/testRequest.jmx
> >>>>>
> >>>>> Schematic View:
> >>>>>
> >>>>>       - http://ubikloadpack.com/demo/schematic/testRequest.html
> >>>>>
> >>>>> Where do the names "handle Cookies", "add Headers", "Virtual Users
> >>>> executing", ... come from? How would extensions (plugins) fit into
> this?
> >>>>
> >>> It currently comes from the xsl file.
> >>> There would be a default display using element name. For now plugins
> would
> >>> have a very basic display.
> >>>
> >>> There would be a need to provide registration for implementations, I
> see
> >>> how to do it for java implementation, but for XSLT, for now I don't.
> >>>
> >> Yes, I think the java implementation would have advantages over the
> >> xsl-only version, especially with respect to extensibility.
> >>
> >>
> >>> Working on this feature (within some consulting work, I was asked to
> >>> document a plan to provide a textual description and ease
> understanding of
> >>> errors (make other users find to what request an assertion was
> related)),
> >>> I
> >>> find that it makes the reading/understanding of a test plan more easy.
> >>> I ran it on previous plans, and I see at least this major benefit.
> >>> Another one (if we move forward) , would be to provide a more source
> >>> version friendly representation of a JMX test plan.
> >>>
> >> Would this lead to something like Vladimir described? A textual
> >> representation of the test plan, possibly with explanations for the
> >> elements, that is diff and reader friendly.
> >>
> >> On the other hand, this reminds me of something. I sometimes dream of
> >> having markers in the test plan tree, that could be used to show
> >>
> >>   * errors that happened during execution of a test plan (linked to/by
> >> samples in result tree)
> >>
> >>   * common misconfigurations like double timers (scoping issues mostly)
> or
> >> beanshell/javascript scripts :)
> >>
> >>   * samplers with long response times
> >>
> >>   * ...
> >>
> >> Felix
> >>
> >>
> >>>
> >>> Felix
> >>>>
> >>>> For now, it's html (I'm using XSLT), but a textual format like YAML
> would
> >>>>> be better for source comparison.
> >>>>>
> >>>>> Regards
> >>>>>
> >>>>> On Mon, Jan 21, 2019 at 11:04 AM Vladimir Sitnikov <
> >>>>> sitnikov.vladimir@gmail.com> wrote:
> >>>>>
> >>>>> Paulo Maia Borges> Why not export to standard HTML?
> >>>>>
> >>>>>> For instance, if jmx is stored in a Git repository (GitLab?
> GitHub?),
> >>>>>> then
> >>>>>> markdown would be much easier to review in diffs than HTML.
> >>>>>> E.g. https://twitter.com/jamiebuilds/status/1085377471057412096
> >>>>>>
> >>>>>> In other words, it could make sense to render markdown
> representation
> >>>>>> on
> >>>>>> save.
> >>>>>> However, it is not clear how tables should be rendered (e.g.
HTTP
> >>>>>> parameters).
> >>>>>>
> >>>>>> Vladimir
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>> <https://www.openstreetmap.org/#map=18/50.69454/3.16455>
> >>>>>
> >>>>>
> >>>>>
>


-- 
Cordialement.
Philippe Mouawad.

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