Christoph Rueger

Related #5112 PR #5116

> more reasonable default, like the number of cores available on the system. I think this is a reasonable suggestion. I tried it out locally, and could not observe a...

based on > We download stuff massively with a new workspace. Since we use a thread per download, and these threads are suspended while waiting for input, we might have...

> So your idea is to use a different, smaller executor pool for HttpClient? Yes, but let's just stay with "different". I shouldn't have mentioned the sizing aspect. Maybe i...

Closing. Feel free to reopen @tjwatson @juergen-albert

> > It should just "be there" but do nothing. > > I think it would be more useful to allow adding repositories to individual (bndrun) files, the whole problem...

> The classpath is controlled completely by the user and nothing is resolved (as far as I know) it actually _look up_ the bundles, What does "controlled by the user"...

Thanks @pkriens , and welcome back :) > I think when we worked on the tags it was implied that we would also apply them to the buildpath. I just...

@pkriens If we introduce a new tag for compilation e.g. `compile` and we already have `resolve` I think we might have created a 7.1.0 backwards compatibility problem situation. Example: -...

I added back the positive tag check "compile" (in addition to negative "nocompile" precedence ) in e467b0b I solved my [backwards compatibility dilemma](https://github.com/bndtools/bnd/pull/6612#issuecomment-2898763421) by now also providing TRANSITION phase: -...