Luca Ferrari

> Could you split it into separate PRs then ? One for each issue - otherwise I can do it as part of the merge. I don't know how to...

Sorry, I didn't realized I had switched to an old base. Now it has been fixed.

I suspect `col_is_pk( name, name, name)` could clash with `col_is_pk( name, name, text)`. How can you distinguish: ` col_is_pk( 'public', 'foo', 'id' )` from ` col_is_pk( 'foo', 'id', 'id is...

I've rebased so far, but at this point I would wait for #329 to be implemented (and I don't know when it will happen, since it is quite challenging). With...

@jesperpedersen can you please give ade40240317bad155dbf1e40866c96257b688b90 a try? I think this is complete now.

See discussion here: . I suspect checking at `barman-wal-restore` time is too late, because PostgreSQL will trigger the WAL recovery and you can only make it failing, and this could...

> 👍 to the idea. > > We just have to figure out what we want in that header... could also be stuff like the entire size of the management...

> An idea could be to be to create a variable length header where each "section" is identified by a unique `char` - like `...cpgagroal-cli\0v1.6.0\0...` with a response of `...Spgagroal\0V1.6.0\0...`...

> Sounds good - a prototype of it will likely give a clearer indication of what is the best solution Ok, I think this will be done in the next...

> is this a possible GSoC 2024 project? I don't think so: this is a quite short (even if not so trivial) change to the communication protocol. @jesperpedersen could add...