By RFC1214:
Before calling a fn, we check that its argument and return types are WF. This check takes place after all higher-ranked lifetimes have been instantiated. Checking the argument types ensures that the implied bounds due to argument types are correct. Checking the return type ensures that the resulting type of the call is WF.
The previous code only checked the trait-ref, which was not enough
in several cases.
As this is a soundness fix, it is a [breaking-change].
Fixes#28609
As described in https://github.com/rust-lang/rfcs/pull/1298, the idea here is to make DefIds independent of changes to the content of other items. They are also *mostly* independent from ordering, so e.g. reordering two functions will change the defids, but it will not change the paths that they expand into (but this is not the case for some things, such as impls).
This is a major refactoring, so I did it in slices. The final commit is in some sense The Big One where most of the work is done. The earlier commits just pave the way by gradually refactoring accesses to the `node` field.
This is a [breaking-change] for plugin authors. The things you need to do to migrate your code are as follows:
1. For local def-ids, rather than do `def_id.node`, call `tcx.map.as_local_node_id(def_id)`.
2. To construct a local def-id, call `tcx.map.local_def_id(node_id)`.
3. Note that you cannot make def-ids for any node, but only for "definitions" -- which include all items, as well as a number of other things, but not e.g. arbitrary expressions.
4. You can get the path to a def-id by calling `tcx.def_path(def_id)`.
One thing that is NOT part of this PR, but which I plan do in a follow-up, is converting uses of the existing `with_path` API to use `def_path`, which is basically the same.
r? @eddyb (or @nrc)
have always been returning None anyway, since it was comparing node-ids
across crates incorrectly -- and remove the now unused map
`extern_const_variants`
paths, and construct paths for all definitions. Also, stop rewriting
DefIds for closures, and instead just load the closure data from
the original def-id, which may be in another crate.
were returned, either the trait or the *self type itself*, were not
particularly representative of what the Def is (a type parameter).
Rewrite paths to handle this case specially, just as they handle the
primitive case specifically. This entire `def_id` codepath is kind of a
mess.
this simplifies the code while reducing the size of libcore.rlib by
3.3 MiB (~1M of which is bloat a separate patch of mine removes
too), while reducing rustc memory usage on small crates by 18MiB.
This also simplifies the code considerably.
I have measured a small, but possibly insignificant, bootstrap performance improvement, and the memory savings grow to about 30M for larger crates (but that is still less as a percentage).
r? @eddyb
This commit swaps the order of linking local native libraries and upstream
native libraries on the linker command line. Detail of bugs this can cause can
be found in #28595, and this change also invalidates the test case that was
added for #12446 which is now considered a bug because the downstream dependency
would need to declare that it depends on the native library somehow.
Closes#28595
[breaking-change]
This commit swaps the order of linking local native libraries and upstream
native libraries on the linker command line. Detail of bugs this can cause can
be found in #28595, and this change also invalidates the test case that was
added for #12446 which is now considered a bug because the downstream dependency
would need to declare that it depends on the native library somehow.
Closes#28595
This is part of #28572, but doesn't complete it. Amongst other things,
this patch:
* Increases consistency in the way feature flags are used with other
docs.
* Removes the ignores, which is nice: we actually had some syntax errors
in the examples 😭.
* Mentions #![no_core]
Realistically, this document used to be in the order of least to most:
nothing, then adding core. But with the changes in RFC 1184, this is
backwards: it now shows stuff that uses core from the beginning. In the
future, I'd like to revamp this to go from 'most to least', but I'd like
to see the discussion in https://github.com/rust-lang/rust/issues/27701
goes before I write more.
this simplifies the code while reducing the size of libcore.rlib by
3.3 MiB (~1M of which is bloat a separate patch of mine removes
too), while reducing rustc memory usage on small crates by 18MiB.
This also simplifies the code considerably.
We don't completely cover documentation tests in the testing chapter,
since we cover them in the documentation chapter. So make sure people
know that.
Fixes#28082
The original blog post referred to examples by their file names, and now
that it's in guide form, there is no file name. So edit the text so that
it makes a bit more sense.
Fixes#28428
The links in the rustdoc for several places in fmt were trying to link to
the std::fmt module but actually linking to std, which was confusing.
While trying to figure out why I noticed that the documentation chapter of
the Rust book has examples that show this same bug (although it doesn't seem
widespread in practice).
r? @steveklabnik
