Hacker News Reader

Comment by zmmmmm | Hacker News Reader

I know probably 95% of HN will see it as a downside but one of my favorite things about DBeaver is that because it is implemented on Eclipse you can install nearly any Eclipse plugin and they work. So I installed Vrapper to edit in Vi-mode, some git tools I like, PlantUML, etc. Really makes it a step above other tools in terms of the power and flexibility you can have with it.

Apart from that, you can install it inside regular Eclipse as a plugin, so then you can have your database window and your ER diagram next to your code. Again, something that really sets it apart from dedicated tools.

View on HN

I love the Eclipse IDE. It's a shame the Eclipse Foundation doesn't treat it better.

Last time I tried to contribute to it, I ended up giving up. They wanted me to sign paperwork simply to report a bug. There was no documentation about how to build the IDE, which is scattered across dozens of repositories. Maybe that has improved, but I doubt it.

They seem to be putting all their eggs into the Theia basket, which is an IDE framework similar to VSCode. Would have been nice to see that energy go towards improving Eclipse instead of reinventing the wheel.

Improving the UI's handling of HiDPI would be nice. One can but dream.

Hi, I am an Eclipse Foundation committer, but I am not committing to the IDE itself.

You don't have to sign the CLA (contributor license agreement) if you want to report a bug, but you have if you want to contribute code, even a small patch for a bug you found. That's a legal requirement to protect the Foundation from contributions that may copy/paste proprietary code and to ensure that you will not in the future change your mind and make claims for your contributions. It's not peculiar to Eclipse Foundation, many open source vendors require contributors to sign a CLA [1]

My understanding is that Theia is currently more of a platform than a complete IDE. It is meant to be usable for building cloud-based IDEs, and you can test a distribution on your workstation, but being it a client/server architecture, you currently have to run a docker container locally to test Theia (or at least this was the case in Q3 2023). They are now focusing on the user interface and plug-in architecture of Theia (which is already capable of running VSCode extension). IMHO, for Theia to become really meaningful the server-side part needs to be improved: currently Theia reuses Eclipse JDT and other plug-ins on the back-end, which is fine for a desktop app, but less convenient for a client/server IDE as you will need a separate server-side container running the Eclipse plug-ins for every active user running the Web-based Theia GUI.

With respect to Eclipse itself, it has dramatically improved over the last few years, since the Foundation started making a release every 3 months. You may like the UI or not, but on my workstation it's now much faster and lighter than it used to be, and it's also faster and lighter than IntelliJ. To me the main issue with Eclipse is that it relies on P2 for plug-in distribution, which makes installing Eclipse add-ons a painful process.

[1] https://en.wikipedia.org/wiki/Contributor_License_Agreement

Hey - don't know what part you contribute to, but thanks for the work whatever you are doing.

Eclipse has been my mainstay for ~20 years now, and while I know lots of people probably more authoritative than me will vehemently disagree, I honestly see it as one of the best pieces of software engineering ever made. The flexibility of the UI, the plugin system, incremental compile, the concept of perspectives and the overall stability delivered given how it is built - a complex ecosystem of cooperating plugins - is absolutely amazing. Now I've reached the point of making my own plugins and contributing to others my respect for it only increased.

It's definitely a complex tool and I think the vast majority of people fall over before they get over the hump of learning to use it. For more experienced users, many people still have in their head the image of Eclipse from 2007 and don't realise that it's been on a path of steady but slow linear improvement for more than a decade - another testament to the underlying engineering.

You might already know, but for others, installing plugins can become a bit less painfull if you make a list of favorites and install those plugins in one go. In order to make favourites list, you have to open an account in eclipse.org and then search in the marketplace website and "star" the required plugins. example -> My favorites URL - https://marketplace.eclipse.org/user/mbhata6m/favorites

I like the Eclipse UI. I just don't like how it behave on high resolution displays on Linux.

I was unable to create an account for the bug tracker without signing a CLA. Maybe that's changed now, but I'm pretty sure that was the case when I last tried.

I wouldn't say that Eclipse has dramatically improved on the last few years. On the contrary. It feels more sluggish to me and it crashes more frequently than before.

The plugin Marketplace in Eclipse is a steaming pile of shite. Installing plugins is an incredibly slow and painful affair. It can take MINUTES to run a search for a plugin, and while it's loading the rest of the UI just freezes. That's a shame.

But given that I can't even figure out how to build the full IDE with its hundrends of plugins, and the CLA nonsense, I certainly won't be investing my time on trying to fix that.

Can't talk for Linux, I am now using Eclipse on Mac.

I used Eclipse on Linux about 8 years ago, and I remember running into frequent issues at that time: the problem was that I was using Ubuntu 16.04, but the SWT widget toolkit maintainers for Linux were Red Hat employees and they were focused mostly on Fedora. For several months SWT was very unstable because the native nature of SWT made it difficult to support it on the different Linux distributions, especially during the transition from GTK2 to GTK3, and considering that Ubuntu was tweaking GTK a lot to achieve its own unique visual style. You can get an idea of the problems the Linux maintainers were having here [1]

At the same time, the release notes of Eclipse clearly stated that that particular version of Ubuntu was not supported. So maybe you are running into the same issue, maybe you are using a Linux distribution that is not amongst the list of officially supported (=tested) Linux distributions.

With respect to market place and plug-ins my experience is different from yours, but having worked on Eclipse plug-ins for years I am certainly a more expert user. As I already said, to me the worst aspect of Eclipse is the p2 package format. It probably solves many problems (e.g. it tries to guarantee that you don't end up with a broken Eclipse instance due to lack of dependencies), but I often think that they could have achieved the same by using plain-old Maven dependencies and BOMs.

[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=492371

Edit: fixed typos

Eclipse started as a "reinvention" of IBM VisualAge for Java and went through at least one major rewrite (transitioning to OSGi). I guess it's the standard pattern for mature code bases.

Would love to learn how you installed these plugins and tested them. Have you written about this somewhere or can you point to some resources that can help people like me, who use dbeaver but can't customize them well.

It's straighforward - it doesn't have the Eclipse Marketplace installed so you have to use direct installation, which just means you have to find the plugin site yourself, which is usually on the web site or github page for the plugin.

So you go to Help => Install New Software and then put eg: the Vrapper update site in:

    http://vrapper.sourceforge.net/update-site/stable

And just go from there.

I've used Vrapper with Eclipse for years, and I've used Dbeaver for years - yet it has never entered my mind before that I could use Vrapper with Dbeaver! Thanks!

> one of my favorite things about DBeaver is that because it is implemented on Eclipse you can install nearly any Eclipse plugin and they work

I like DataGrip for a very similar reason, relative to IDEA plugins.

Potentially good info to know for someone using Eclipse.

Why start your comment by saying "I know probably 95% of HN will see it as a downside but"?

And this is irrespective of whether that 95% is right or not.

Google "father son and donkey story".

And re: that 95%:

See normal distribution aka bell curve.

And echo chamber and hive mind.

All of the above imply the same main point, just in different ways.

Unfortunately I pretty much have only ever seen negative comments about Eclipse on HN. It seems immensely unpopular and often the minute you mention it you'll get a bunch of often derisive replies. At this point I've had to come to terms with the fact that I'm the odd one out and own it, and that is how I do it. No disrepect meant, more acknowledgement of my own exceptionalism.

Interesting approach.

Sort of like defence against possible downvoters.

If OP went along the "father, son, and donkey story," then the OP won't have posted it at all.

No, they could still have chosen to post it, even if they went along with that story.

My point was that there is no need to prepare a defence against downvoters, that too, up front. Not needed at all. Why is their opinion important to the OP? They can have their opinions, he can have his opinion of the same matter. What's the problem? The problem is only if one thinks that HN karma is important. It's totally unimportant.

>he can have his opinion

he or she (or they), sorry.

Karma is unimportant, but downvotes fade away the comment. There is little point in making a comment that will disappear anyway, since no one will see it.

>Karma is unimportant, but downvotes fade away the comment.

So what? IMO, HN comments are also unimportant.

In fact, HN is also unimportant. I said that here on HN recently.

Jeez. Does anyone think HN is important in the real world?

In the HN world, HN may be important (and even then, only to some), but then it is an echo chamber.

>There is little point in making a comment that will disappear anyway, since no one will see it.

Maybe. Maybe not. Do you rate the importance or value of your comments by the number of people who see them?

>There is little point in making a comment that will disappear anyway, since no one will see it.

So what? So comment anyway, without bothering about it disappearing, just to express yourself, or refrain from commenting if the disappearing of your comment is so important to you.

That makes sense.

Why are you here, though?

Not sure of your meaning, but maybe you're confusing unimportant with uninteresting? HN may be unimportant, but it is (sometimes) interesting, because of those very comments. Go figure (pun intended).

>That makes sense.

First say which of my above statements you refer to.

Until then I cannot meaningfully reply to your question:

>Why are you here, though?

Also, why should I not be here? You have any problem with it?

See what I did there?

> First say which of my above statements you refer to.

Nothing in particular.

> Also, why should I not be here? You have any problem with it?

Not at all.

You might be reading too much into my comment, though.

Dude:

>>First say which of my above statements you refer to.

>Nothing in particular.

>You might be reading too much into my comment, though.

I cannot be reading too much into your comment if you say that you refer to nothing in particular (out of my statements).

QED.

Be logical, like Spock:

https://en.m.wikipedia.org/wiki/Leonard_Nimoy

kthxbye :)

> Karma is unimportant, but downvotes fade away the comment.

Preemptive defense against downvotes is a good way to attract them, so that's another reason not to do it.

>so that's another reason not to do it.

Unless one doesn't care about downvotes, because, you know, one is not a slave to convention.

Reply on news.ycombinator.com