Cleanup and refactor FnCtxt::report_no_match_method_error #148652
Conversation
rustbot
added
S-waiting-on-review
|
r? @wesleywiser rustbot has assigned @wesleywiser. Use |
|
r? @lcnr |
Currently this method is quiet long and complex, this commit improves its readability by adding sub-fns
rperier
force-pushed
the
report_no_match_method_error-refactoring
branch
from
9a8910e to
177bb33
Compare
|
The job Click to see the possible cause of the failure (guessed by this bot) |
lcnr
changed the title
|
For changes which move a bunch of code around it's really helpful to split it into individual commits, e.g. if you have "move things into a |
| Some(sugg_span), | ||
| ); | ||
| } | ||
| custom_span_label |
There was a problem hiding this comment.
return true here and do a fast return if static_candidates.is_empty(). Or probably even only call this function is static_candidates aren't empty
| item_kind: &str, | ||
| mode: Mode, | ||
| source: SelfSource<'tcx>, | ||
| unsatisfied_predicates: &Vec<( |
There was a problem hiding this comment.
&Vec should pretty much always be a slice?
| ) { | ||
| let mut find_candidate_for_method = false; | ||
|
|
||
| if should_label_not_found && !custom_span_label { |
There was a problem hiding this comment.
do this in the caller and only have a single bool argument?
| if rcvr_ty.references_error() { | ||
| err.downgrade_to_delayed_bug(); | ||
| } |
There was a problem hiding this comment.
odd, why does this happen here and not at the start of the fn/before we even start building the error? Can you move this up as far as possible
| let mut static_candidates = no_match_data.static_candidates.clone(); | ||
|
|
||
| // `static_candidates` may have same candidates appended by | ||
| // inherent and extension, which may result in incorrect | ||
| // diagnostic. | ||
| static_candidates.dedup(); |
There was a problem hiding this comment.
why does this happen outside of suggest_static_method_candidates? do we use the deduplicated static_candidates anywhere else?
| let mut restrict_type_params = false; | ||
| let mut suggested_derive = false; | ||
| let mut unsatisfied_bounds = false; |
There was a problem hiding this comment.
these 3 bools are very ☠️ would be nice to look into how to make them unnecessary or somehow unify their intended impact
| ) { | ||
| return guar; | ||
| } | ||
| span = item_ident.span; |
There was a problem hiding this comment.
that's weird. why? and could we just use item_ident.span directly below this point?
5 participants
As discussed on zulip with @lcnr, this is a proposal for refactorizing this method.
See #t-compiler/help > (PR #144674) Merge multiple suggestions into a single @ 💬