Split off from #62855
This PR deerializes the declaration `Span` and attributes for all
procedural macros from their underlying function definitions.
This allows Rustdoc to properly render doc comments
and source links when inlining procedural macros across crates
Avoid the tricky scheme with callbacks and keep the invocation parent data where it logically belongs - in `Definitions`.
This also allows to create `InvocationData` entries in resolve when the data is actually ready, and remove cells and "uninitialized" variants from it.
Replaced linear token counting macros with optimized implementation
There are currently two distinct token-counting macros in the source. Both implement the trivial algorithm, with linear complexity. They may or may not be adequate for their use case, but considering that other people are probably going to copy and paste them whenever they need a token-counting macro, I replaced them with an optimized implementation with logarithmic complexity.
This commit makes two changes - separating the `NodeId` that identifies
an enum variant from the `NodeId` that identifies the variant's
constructor; and no longer creating a `NodeId` for `Struct`-style enum
variants and structs.
Separation of the variant id and variant constructor id will allow the
rest of RFC 2008 to be implemented by lowering the visibility of the
variant's constructor without lowering the visbility of the variant
itself.
No longer creating a `NodeId` for `Struct`-style enum variants and
structs mostly simplifies logic as previously this `NodeId` wasn't used.
There were various cases where the `NodeId` wouldn't be used unless
there was an unit or tuple struct or enum variant but not all uses of
this `NodeId` had that condition, by removing this `NodeId`, this must
be explicitly dealt with. This change mostly applied cleanly, but there
were one or two cases in name resolution and one case in type check
where the existing logic required a id for `Struct`-style enum variants
and structs.
Before this two macros with same name would be indistinguishable inside a
`FileName`. This caused a bug in incremental compilation (see #53097) since
two different macros would map out to the same `StableFilemapId`.
Fixes#53097.
At least the incremental compilation code, and a few other places in the
compiler, require the CrateMetadata for a loaded target crate to contain a
valid DefIdTable for the DefIds in the target.
Previously, the CrateMetadata for a proc macro contained the crate's
"host" DefIdTable, which is of course incompatible with the "target"
DefIdTable, causing ICEs. This creates a DefIdTable that properly refers
to the "proc macro" DefIds.
Fixes#49482.
Michael Woerister pointed out that `hir_to_node_id` (introduced in
August 2017's 28ddd7a4e) supersedes the functionality of
`find_node_for_hir_id` (just a hash lookup compared to that linear
search).
