Replace XzLibLzmaDevelMigrator with MiniReplacement
#3851
Conversation
Introduce a `MiniReplacement` migrator that is a minimigrator based on `Replacement`, and replace the `XzLibLzmaDevelMigrator` with it. Extend to v1 recipes while at it.
| # TODO: do we want to limit to host requirements, like the split | ||
| # minimigrators did? |
There was a problem hiding this comment.
We should make this an option and then turn it on as needed to match the current behavior. For ABI migrations, we only ever want to edit host.
| *, | ||
| old_pkg: "PackageName", | ||
| new_pkg: "PackageName", | ||
| requirement_types: tuple[str] = ("host",), |
There was a problem hiding this comment.
Oh I am sorry I was mistaken. The old "xz" one did everything in the text file as long as it matched xz\n. So I think actually what we want is none of the above and something much more general. We should inspect the other migrators to see what they look like too.
There was a problem hiding this comment.
Well, yes and no. It replaced everything but filter()-ed on host requirements. In that sense, the new minimigrator does the same thing. Ofc, I'm open to changing the behavior to something "saner".
There was a problem hiding this comment.
Right. My concern is us missing things like ignored run exports in build/ignore_run_exports_from, etc. The current migrator would catch that and we need to replicate that kind of behavior. I don't have an exhaustive list of where say xz might appear, which is why something more generic would be better.
Description:
Introduce a
MiniReplacementmigrator that is a minimigrator based onReplacement, and replace theXzLibLzmaDevelMigratorwith it. Extend to v1 recipes while at it.Checklist:
Cross-refs, links to issues, etc:
Discussed in #3360
@beckermr, could you do a quick initial review before I switch some more minimigrators over?