Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [pdt-dev] Pushing project forward


I don’t want any “hold-on” step. I can off course imagine one cycle (3 months) delay between releases, but shouldn’t be bigger mostly because I’im using PDT for my daily work ;)

During previous releases we successfully replace a lot of ancient code (for example Indexer, Reconciler madness) , improve performance and introduce a lot of new features.
DLTK had a lot of amazing stuff that move PDT forward in past, but this was during PHP 4 and 5. PDT 5.3 introduces non-java namespace model, and even this wasn’t implemented correctly due model limitation. 
Later PHP releases only escalate problems (traits, function/constant imports). As the result in most modern project, PDT have to dive  into AST to check what is going on…

I prefer to introduce new model, and adopt interesting parts of DLTK (search engine, indexing) to IT. Most important things (Inference Engine and Code Assist) are fortunately mostly ready for that. 

I think we should be able to move all current bugs into GitHub / GitLab.

On 27 January 2020 at 18:35:47, thierry blind (thierryblind@xxxxxxx) wrote:

Hi everybody,
thank you Michele and Basil Mohamed Gohar, your point of view are very accurate to me.

My first concern is, probably like many open-source developers, that the projects I contribute are perene.
I've seen too many (little and big) projects dying, by lack of interest of the community or by the main developers themselves.
Well yes, sometimes it's for more basic reasons, developers also need to eat and earn money, and their job can take all their time and energy.
Also with time, developers change their occupations to try new challenges.
If I look at eclipse itself now, I find it myself less "sexy" than other newer technologies, based on nodejs for example (I was thinking at Visual Studio Code).

But PDT is a great product and has still a lot of potential.
PDT is daily used by myself, by my developer colleagues and by other people I know and they are happy with it.
Like Dawid said, PDT code changed a lot the 5 last years, and in the right direction.
It's much more stable now that when I started to contribute to PDT, a lot of work was done by a little team, and Dawid contributed A LOT, because he has a good global view/skill of the eclipse plateform.

PDT has a lot of feature and I think it's really worth (at least for me) to put some of my free time to make it even better.
But the more I do, the more I see that 2 active developers are far from being enough for such a big project, for many logical reasons I won't enumerate here (one being lack of free time).
If we migrate to github, to have more visibility, we will maybe be overwhelmed by feature or evolution requests, and maybe also by bug reports that are already on bugzilla.
So migrating to github? Yes for sure. But without new developers, I feel it will be a pain with our actual capacities and it will add an additional workload.
And that's my real concern about PDT. We have to stay realistic about our real capacities and reactivity to answer user requests, or we will generate more frustration and delays. In the 500 opened bugs, there are a lot of enhancement requests, and many are justified.

For the 4 technical points, I'm agree by removing old stuff and adopt more adequate technologies, even if we need to rewrite the PHP model and other things.
We have a lot of code that needs to (or could) be revised to be more efficient, but with the proposed changes, it's like developing a new product from the scratch. I feel that many PDT code parts have reached their technical "limit".

So, what is the best solution? Rewrite big parts one after the other, probably breaking a lot of things and delaying new PDT releases?
Or puting actual PDT in maintenance mode and rewriting in parallel a whole new PDT product?
Keep in mind that we are only 2 developers yet and that Dawid has a lot more PDT/eclipse technical background then me 😉


De : pdt-dev-bounces@xxxxxxxxxxx <pdt-dev-bounces@xxxxxxxxxxx> de la part de Michele Locati <michele@xxxxxxxxx>
Envoyé : lundi 27 janvier 2020 14:52
À : PDT Developers <pdt-dev@xxxxxxxxxxx>
Objet : Re: [pdt-dev] Pushing project forward
Hi Dawid,

sorry for not being able to contribute to this project.
In addition to a chronic lack of spare time, for sure the knowledge of Bugzilla/Gerrit/GIT/Jenkins is a bit of a barrier for newcomers... I committed a couple of changes a while back, and if I had to do that again today it'd take me an hour or two just to remember how to do that.
If compared to the standard GitHub flow (clone + create a pull request), contributing to PDT it very over-complicated IMHO.

Furthermore, being on GitHub gives more visibility to the project, and integrating with GitHub Actions to do tests is a real pleasure.

So, I'm 100% for moving to GitHub...


Il giorno lun 27 gen 2020 alle ore 15:03 Dawid Pakuła <zulus@xxxxxxxxx> ha scritto:

Couple days I opened thread for Handly migration. Despite to final results I looking for way to get involved more people into project.

For now we have two active comitters, Me and Thierry. We have couple supporters on bugzilla (for ex. Filipus and Michele). Due a this I’m afraid that sooner or later project will be dead or at stagnated.
Our PHP package is quite popular, and we have stable number of downloads: ~30 000 for EPP + ~10 000 from marketplace. Some distributions like Ubuntu have own packages so don’t know everything.

Our code base evolved a lot since last 5 years, and now is much easier to implement new features or improve existing, but we still have incredible mess :P

So, how to push project forward?

1. Validation and PHP model :  as I wrote earlier I think we should abandon some DLTK features and implement own model and inferencer. Without this upcoming PHP 8 will be extremely hard to implement in efficient way.
2. _javascript_ integration : JSDT is dead, we should add Wild Web Developer features
3. We have over 500 bugs, and they are still relevant
4. Nuclear option : switch to Generic Text Editor without WTP

Project management:
I liked Eclipse Bugzilla/Gerrit/GIT/Jenkins infrastructure but most of our users probably prefer GitHub or GitLab. I think we should consider switching to GitHub or upcoming Eclipse GitLab instance. 

Please say what you think? Commiters, contributors, users, adopters? 

pdt-dev mailing list
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
pdt-dev mailing list
To change your delivery options, retrieve your password, or unsubscribe from this list, visit

Back to the top