David E. Wheeler

Yeah, totally, that would be very cool, thanks!

Yikes! That's no fun. Looks like that query gets called for each change to be reverted, to make sure no other changes depend on it: https://github.com/sqitchers/sqitch/blob/a2c920545cdba64367c88b77a6268b0fc503f0d3/lib/App/Sqitch/Engine.pm#L626-L645 The query is defined...

That still would be executed multiple times, though, right? Ideally, we could design a single query to take an array of change IDs, and return their dependencies — but only...

Putting this off for now. I think it will be complicated. Lots of code touches the URIs, an in particular: - Deploy would need to deploy one change at a...

No, not recently, and I haven't had the chance to work on Sqitch for a while. Contributions warmly welcomed, though.

Yeah, that would be the way to update it if you wanted to change it, assuming you haven't deployed it to to a lot of locations that would be a...

Oh, right, the URI is included in the hash for changes, isn't it? You'd have to use a [`sqitch plan`](https://sqitch.org/docs/manual/sqitch-plan/) format to get a list of changes and their SHA1...

Maybe do this: ``` sqitch plan -f 'format:%H %n' > before.txt ``` Then change the URI in the plan file and run it again: ``` sqitch plan -f 'format:%H %n'...

It's not just the Sqitch config file, but also configuration files used by each engine.

Well, we could easily teach to favor `$XDG_CONFIG_HOME` over `$HOME` when looking for files, but I don't know whether and how well each engine supports that environment variable when reading...