You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jmeter.apache.org by Vladimir Sitnikov <si...@gmail.com> on 2019/10/05 18:58:30 UTC
Darcula theme vs tree view (the tree of components)
Hi,
TL;DR: what if we copy Darcula sources into JMeter repository and fix it a
bit?
At the end of the day, Darcula is AL2.
+1: Of course, let's copy it, and drop when upstream publishes a new release
+0.5: Ok, let's copy it. And drop when upstream publishes a new release
0: JMeter UX is perfect
-0: ?! Copy-paste development is forbidden
-1: Do not do that because ...
I've got a report that Darcula theme makes it hard to understand the
component tree.
The user seems to like the dark theme, however, the UX is bad, so they
revert to the classic one.
The issue is nesting is barely visible because of the very small offset in
the tree.
A sample screenshot:
https://jmeter.apache.org/images/screenshots/changes/5.0/jmeter_5_search_feature.png
The items "RA_Redirected" and "HR_NoAutoRedirect" look as if they are
placed under the same parent.
Remember that the picture represents a very basic test plan, and for more
involved cases it would be extremely confusing.
It seems to be a regression when compared with the old theme:
https://jmeter.apache.org/images/screenshots/http-config/http-config-example.png
We ship with Darcula by default, so it is sad to have a bad-looking tree
there.
I've tracked the issue, and the offset is hard-coded in Darcula repository:
https://github.com/bulenkov/Darcula/blob/49774425c392c36d926b8a4518a3616a8cf735e2/src/com/bulenkov/darcula/ui/DarculaTreeUI.java#L60
However, the author says he won't fix current Darcula, as "the upstream"
(==Darcula that is present in IntelliJ sources) is way ahead.
In other words, the author suggests he (or somebody else, because, well, it
is OSS) would factor out "IntelliJ Darcula" into its own submodule, then he
could grab it and publish it as a separate artifact.
See https://github.com/bulenkov/Darcula/issues/56#issuecomment-526942616
However, I don't think it would happen soon taking the number of Darcula
releases into account (1 release per 6 years).
Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by Vladimir Sitnikov <si...@gmail.com>.
>What is the licence?
Apache 2.0, NOTICE is absent.
Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by sebb <se...@gmail.com>.
What is the licence?
On Sun, 6 Oct 2019 at 15:28, Vladimir Sitnikov
<si...@gmail.com> wrote:
>
> >so do I
>
> Ok, then I stop commenting in this thread, and just copy the sources.
> We are losing time for no reason.
>
> Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by Vladimir Sitnikov <si...@gmail.com>.
>so do I
Ok, then I stop commenting in this thread, and just copy the sources.
We are losing time for no reason.
Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by Philippe Mouawad <pm...@apache.org>.
On Sunday, October 6, 2019, Vladimir Sitnikov <si...@gmail.com>
wrote:
> >why not help darcula doing what Bulenkov has suggested ?
>
> Do you have free cycles to do that?
No but what is the difference with importing code and fixing vs forking,
fixing and pulling request ?
We would use the fork
> Go ahead.
>
> I'm afraid my time is limited, so I have to pick which projects to work on.
so do I
>
> Including Darcula to JMeter would fix a critical end-user issue at very
> minimal support cost for the development team.
>
> Vladimir
>
Re: Darcula theme vs tree view (the tree of components)
Posted by Vladimir Sitnikov <si...@gmail.com>.
>why not help darcula doing what Bulenkov has suggested ?
Do you have free cycles to do that?
Go ahead.
I'm afraid my time is limited, so I have to pick which projects to work on.
Including Darcula to JMeter would fix a critical end-user issue at very
minimal support cost for the development team.
Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by Philippe Mouawad <pm...@apache.org>.
On Sunday, October 6, 2019, Vladimir Sitnikov <si...@gmail.com>
wrote:
> >as we are committing ourselves way more to the maintenance of
> >that software than to just updating the dependency
>
> In fact, Darcula theme is already distributed, so the maintenance is the
> same.
>
> The next version of Darcula would likely be a complete rewrite, so the
> upgrade will anyway be painful.
I don’t think upgrade would be painful, we have very few classes that
depend on it AFAIR.
> As the next "official" version appears, we could drop the relevant sources
> and use the dependency.
>
> >* Make an external fork of darcula and use that one (Would have the same
> >problems with copying the sources in my opinion)
>
> What is the point?
> Note: it would be extremely hard to make coordinated changes.
> If we just copy the source, then we could fix it and immediately see how
> that plays with JMeter.
>
> If someone forks Darcula, then we can't really use it in JMeter until the
> fork releases. Two repositories make things much harder.
why not help darcula doing what Bulenkov has suggested ?
>
> Vladimir
>
Re: Darcula theme vs tree view (the tree of components)
Posted by Vladimir Sitnikov <si...@gmail.com>.
>as we are committing ourselves way more to the maintenance of
>that software than to just updating the dependency
In fact, Darcula theme is already distributed, so the maintenance is the
same.
The next version of Darcula would likely be a complete rewrite, so the
upgrade will anyway be painful.
As the next "official" version appears, we could drop the relevant sources
and use the dependency.
>* Make an external fork of darcula and use that one (Would have the same
>problems with copying the sources in my opinion)
What is the point?
Note: it would be extremely hard to make coordinated changes.
If we just copy the source, then we could fix it and immediately see how
that plays with JMeter.
If someone forks Darcula, then we can't really use it in JMeter until the
fork releases. Two repositories make things much harder.
Vladimir
Re: Darcula theme vs tree view (the tree of components)
Posted by Felix Schumacher <fe...@internetallee.de>.
Am 05.10.19 um 20:58 schrieb Vladimir Sitnikov:
> Hi,
>
> TL;DR: what if we copy Darcula sources into JMeter repository and fix it a
> bit?
> At the end of the day, Darcula is AL2.
>
> +1: Of course, let's copy it, and drop when upstream publishes a new release
> +0.5: Ok, let's copy it. And drop when upstream publishes a new release
> 0: JMeter UX is perfect
> -0: ?! Copy-paste development is forbidden
> -1: Do not do that because ...
I think the +-X answers are chosen quite suggestive, but that is a
completely different issue.
I am not very comfortable with including the source of darcula in our
repo, as we are committing ourselves way more to the maintenance of
that software than to just updating the dependency. I am unsure whether
we have the expertise to maintain it in the long run.
Are there any other alternatives to copying the sources into our repo?
I could think of:
* Make an external fork of darcula and use that one (Would have the same
problems with copying the sources in my opinion)
* Drop darcula (as it seems to be not well enough maintained for our means)
Do we have enough people in our community, that are willing and capable
of maintaining it?
For me it is a -0.5 at the moment for fear of long term burden on the
team and my missing expertise in GUI programming.
Felix
>
>
> I've got a report that Darcula theme makes it hard to understand the
> component tree.
> The user seems to like the dark theme, however, the UX is bad, so they
> revert to the classic one.
> The issue is nesting is barely visible because of the very small offset in
> the tree.
>
> A sample screenshot:
> https://jmeter.apache.org/images/screenshots/changes/5.0/jmeter_5_search_feature.png
>
> The items "RA_Redirected" and "HR_NoAutoRedirect" look as if they are
> placed under the same parent.
> Remember that the picture represents a very basic test plan, and for more
> involved cases it would be extremely confusing.
>
> It seems to be a regression when compared with the old theme:
> https://jmeter.apache.org/images/screenshots/http-config/http-config-example.png
>
> We ship with Darcula by default, so it is sad to have a bad-looking tree
> there.
>
> I've tracked the issue, and the offset is hard-coded in Darcula repository:
> https://github.com/bulenkov/Darcula/blob/49774425c392c36d926b8a4518a3616a8cf735e2/src/com/bulenkov/darcula/ui/DarculaTreeUI.java#L60
>
> However, the author says he won't fix current Darcula, as "the upstream"
> (==Darcula that is present in IntelliJ sources) is way ahead.
> In other words, the author suggests he (or somebody else, because, well, it
> is OSS) would factor out "IntelliJ Darcula" into its own submodule, then he
> could grab it and publish it as a separate artifact.
> See https://github.com/bulenkov/Darcula/issues/56#issuecomment-526942616
> However, I don't think it would happen soon taking the number of Darcula
> releases into account (1 release per 6 years).
>
>
> Vladimir
>