eclipse.platform.releng.aggregator icon indicating copy to clipboard operation
eclipse.platform.releng.aggregator copied to clipboard

Move org.eclipse.jdt.doc.user to JDT repository

Open laeubi opened this issue 2 years ago • 9 comments

Looking at https://github.com/eclipse-platform/eclipse.platform.releng.aggregator/blob/master/eclipse.platform.common/bundles/org.eclipse.jdt.doc.user/pom.xml this seems to require nothing special from aggregator build itself.

So it seems suitable to be maintained at the jdt project itself.

WDYT @akurtakov @iloveeclipse @mickaelistria can you probabbyl help with creating a PR with full history if we see if/how this can work out at the JDT side? We then probably need to update some references in aggregator afterwards.

laeubi avatar Oct 30 '23 10:10 laeubi

jdt.doc.isv as well. But I see content in parent pom at https://github.com/eclipse-platform/eclipse.platform.releng.aggregator/blob/master/eclipse.platform.common/pom.xml which seems to be common to all docs bundles. Do we want to un-factorize this part? (I personally don't have a strong opinion here).

mickaelistria avatar Oct 30 '23 10:10 mickaelistria

jdt.doc.isv as well.

That is not that easy but I'm working on it, org.eclipse.jdt.doc.user on the other hand seems like a low hanging fruit if I do not overlook anything?

laeubi avatar Oct 30 '23 10:10 laeubi

IMO, it's better to migrate do both or none here. Migrating just one would make things a bit inconsistent and I don't think it would actually help overall.

mickaelistria avatar Oct 30 '23 10:10 mickaelistria

Migrating just one would make things a bit inconsistent and I don't think it would actually help overall.

Why? as shown org.eclipse.jdt.doc.user seem not to relate to anything else (aka self-contained), while jdt.doc.isv has dependencies to literally everything in the aggregator (see org.eclipse.jdt.doc.isv/cbi_basedirs.properties).

But I see content in parent pom at https://github.com/eclipse-platform/eclipse.platform.releng.aggregator/blob/master/eclipse.platform.common/pom.xml which seems to be common to all docs bundles. Do we want to un-factorize this part?

I think we disabled indexing altogether already so this is just dead code.

laeubi avatar Oct 30 '23 10:10 laeubi

Why?

Imagine some contributor saying "I want to contribute to JDT doc". The current state is to just go to the right repo and edit it. If we split, then the same "I want to contribute to JDT doc" now brings additional questions before finding the right thing to do. So I don't think it makes things simpler is one more step is necessary before being able to produce value.

mickaelistria avatar Oct 30 '23 10:10 mickaelistria

The current state is to just go to the right repo and edit it.

You seem to asume that the user already know what is "the right repo", currently always the user needs to head over to a completeley different alien place, while with moving the user doc at laest that can be found already next to the code.

But of course this was jsut an idea of something we can act on now than sometime in the future (if ever).

laeubi avatar Oct 30 '23 10:10 laeubi

You seem to asume that the user already know what is "the right repo"

No. I also think about reviewers (us) who, if a user wants to contribute to the doc, will have to constantly ask them "What do you want to contribute? Is this final user documentation or API/framework documentation?" and start a longer conversation while the current state is just to answer 1 link. Moreover, some changes might affect both user and isv at the same time and one might prefer 1 PR for both rather than 2 PRs in 2 distinct repos. That's why I think that both should move when they're ready; but moving only user doc now seems like a good intention, but with no so positive outcome.

mickaelistria avatar Oct 30 '23 11:10 mickaelistria

Should the PDE docs also be moved to the PDE repo?

vogella avatar Nov 01 '23 17:11 vogella

@vogella see

  • https://github.com/eclipse-platform/eclipse.platform.releng.aggregator/issues/1497

laeubi avatar Nov 01 '23 19:11 laeubi