[#1098] Detect unused record fields #1099
Conversation
robertoaloi
force-pushed
the
1098-detect-unused-record-fields
branch
3 times, most recently
from
e6c43c9
to
532a154
Compare
robertoaloi
changed the title
robertoaloi
force-pushed
the
1098-detect-unused-record-fields
branch
from
532a154
to
1cb142e
Compare
robertoaloi
force-pushed
the
1098-detect-unused-record-fields
branch
from
1cb142e
to
e43534b
Compare
| {RecordName, RecordField} -> | ||
| els_utils:to_binary( | ||
| lists:flatten(io_lib:format("~p/~p", [RecordName, RecordField]))); | ||
| Id -> atom_to_binary(Id, utf8) |
There was a problem hiding this comment.
in which case can the Id be an atom?
There was a problem hiding this comment.
That's copy-paste from the Unused Macro diagnostics backend, cleaning.
| case POIId of | ||
| {RecordName, RecordField} -> | ||
| els_utils:to_binary( | ||
| lists:flatten(io_lib:format("~p/~p", [RecordName, RecordField]))); |
There was a problem hiding this comment.
nitpick: what about ~p.~p or #~p.~p? (looks more recody syntax)
| FullName = | ||
| case POIId of | ||
| {RecordName, RecordField} -> | ||
| els_utils:to_binary( |
There was a problem hiding this comment.
els_utils:to_binary( supports deep lists (unicode:chardata()) so no need to flatten. On the other hand should we use ~tp instead of ~p to properly print unicode atoms? (I am always confused with unicode, maybe a new atom_to_label function in els_utils could have it right once and that could be used in the whole code base)
There was a problem hiding this comment.
I like that approach. So, I will keep it for now and handle this as a follow up (feel free to contribute on that).
| find_unused_record_fields(Document) -> | ||
| Definitions = els_dt_document:pois(Document, [record_def_field]), | ||
| Usages = els_dt_document:pois(Document, [record_field]), | ||
| UsagesIds = [Id || #{id := Id} <- Usages], |
There was a problem hiding this comment.
optimisation: UsagesIds could be an usort-ed list
Description