I'll only consider part where my reply can be useful and possibly educational.
3.2) also the code review, but i have no idea how it was done in the
past or how to approach it;
I'm unaware that GS has a 'code review' tool. Perhaps sometimes people do it via Savannah?
Related to gnustep-on-android, I once reviewed and merged portions of a relatively small patch dump delivered via email. It isn't an experience I want to repeat.
mozilla/chatzilla review code by submitting
patch files to bugzilla and typing comments there and i was using this
approach and it worked just fine, as would savannah's patch section or
email for that basic level of functionality. if more is needed i would
like to know what it is - this is not an attempt to say that someone is
stupid for suggesting that this is important, i just do not understand
the problem clearly in my mind
Once you see how pleasant it is to leave review comments _inline_, you don't go back.
That's just a start. Diffing "revision 3 of the change" to "revision 7 of the change" when confirming that comments left on revision 3 have been correctly addressed is also very pleasant. ("So, you marked this thread originating in revision 3 as 'done', but I can see that you didn't fix this style issue even in revision 7. Can you revisit it?") This is much more useful than looking at diff against base, or diffing diffs.
This is offered by Gerrit, and I have successfully mentioned it multiple times on these mailing lists.
As a bonus that we have, in principle, agreed not to use: Gerrit also offers 'enforced approval': you can't submit until someone reviews and agrees that the code is in a submittable state. You can also enforce verification by a CI system.
4) discusisng (3) and the like is important, but i see little discussion
about it, and it is astonishing
Discussing things like (3) is hard via email. Usually a team member would research options, come with a recommendation, present it, and possibly present alternatives -- in a live meeting.
We don't have frequent enough live meetings, nor video conferences.
In Dublin, we have, in principle, agreed that we would have rare, but regular, 'real-time' meetings of contributors. Nothing has been scheduled thus far.