Looks good to me, but I want someone else to put eyes on it before it gets pushed.

The only thing I'm curious about really is why we have extrastorage and usualstorge. Can we get away with only one of those choices instead of having more options?

I'm not thrilled with this - it's a lot of code, and I think there are some simpler options that haven't been explored. For example, clients could use the logexchange infrastructure to prime the sampling process and do a lot better.

Seems fine, but probably file a bug against mypy on this.

Did you run tests on this patch? IIRC there's a test that'll fail ;)

I agree with marmoute on all points.

Queued with updated log message. Thanks!

I like the feature, I'll go ahead and queue it with the str/bytestr thing fixed in flight. Thanks!

I agree with Martin's request for changes here.

needs rebased

I still think we should fix the cwd for fixers, but it seems Martin and I agree that repo-root is more consistent. Would you be willing to reroll with repo-root instead of cwd?

I'm landing this, but in the future I suspect we should make sure we just have a mix of hg graft --continue and hg graft rather than run the whole test twice, as it'll severely bloat the already-slow testsuite to do that for many tests.

(Martin, this looks fine to me, if you've got no objections let's include it in 5.1)

I know @martinvonz has been looking at files in changelogs recently. Do you have an opinion on this? It looks fine to me.

In D6597#97228, @Alphare wrote:

In D6597#97177, @durin42 wrote:

LGTM, but needs rebased (sorry for slow response)

@durin42 I've rebased it on hg-committed, but nothing's changed. Am I missing something?

LGTM, but needs rebased (sorry for slow response)

Per your note that this will probably make other localizations sad, let's just give up on this for now?

Why? Before it was sorted, and now it's not.

In D6562#95823, @spectral wrote:

In D6562#95775, @durin42 wrote:

I like this, but could I get you to add a patch 3 that introduces a develwarn for this?

Done, and it found a couple other cases, but I'm not sure I like it (D6573) - non-English users might not be able to enable the devel warnings without getting a bunch of spam. I believe by the time we get to ui.prompt() we've already done the _() translation and it's in the target language? Do we have any idea if anyone enables developer warnings outside of tests (which would typically be written in English, at least for core).
Unfortunately, I don't think it's something we could put in check-style.py either... we won't know that these are going to a prompt. Anyway, we can continue discussion on this on D6573.

I like this, but could I get you to add a patch 3 that introduces a develwarn for this?

Note that I mailed a version to the list for stable. This doesn't quite cleanly apply there, so I made the change twice so I could write the next change without waiting for a merge.