ignore irrelevant symbols when pinpointing undefined variable lints #1915
Conversation
Codecov Report
@@ Coverage Diff @@
## main #1915 +/- ##
=======================================
Coverage 98.95% 98.95%
=======================================
Files 113 113
Lines 4953 4966 +13
=======================================
+ Hits 4901 4914 +13
Misses 52 52
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
R/object_usage_linter.R
Outdated
| descendant::SYMBOL | ||
| | descendant::SYMBOL_FUNCTION_CALL | ||
| descendant::SYMBOL[not(ancestor::expr[OP-TILDE])] | ||
| | descendant::SYMBOL_FUNCTION_CALL[not(ancestor::expr[OP-TILDE])] | ||
| | descendant::SPECIAL |
There was a problem hiding this comment.
Could this also appear in a formula?
There was a problem hiding this comment.
I think any infix operator used in a formula has to be defined elsewhere, so no. trying to thing of an example...
foo <- function() {
lm(y ~ X %**% Z, data = X %**% Y)
}If checkUsage is dinging %**%, it should also be dinged in the formula.
The only thing I can think of is where the operator has some special definition inside the formula, which doesn't happen in base since only %in% has special interpretation:
https://stat.ethz.ch/R-manual/R-devel/library/stats/html/formula.html
I guess it's conceivable for other downstreams to define special interpretation methods for formulas, but because checkUsage() ignores formulas entirely (IIUC), it would have to be that both (1) %**% has a special interpretation and (2) it's being used incorrectly outside of a formula.
That seems pretty out there...
MichaelChirico
changed the title
|
gentle ping @AshesITR |
Part of #1914
I'm not a fan of the
descendant::*[ancestor::*]XPath, but I'm not sure it can be avoided.