Pylint only supplies the start position of markers, but not the end. Always extending the marker until the end of the line leads to undesirable highlighting in editors, for example:
Starting an "unused argument" diagnostic at unused_arg and extending It until EOL would also highlight used_arg, even though the diagnostic doesn't apply there.
With this commit, the end position sent to the editor is automatically determined by looking for an identifier at the indicated start position. If there is no valid identifier at that position, the entire line is highlighted as before.
Note that in some cases, Pylint doesn't supply the start position of the token of interest, but instead marks the first non-whitespace character. I noticed this on import statements and except clauses, where the import (or from) and the except keywords then are what gets highlighted, instead of the module or variable in question. However, I think these are only minor issues in comparison to the original issue with the example above.
Pylint only supplies the start position of markers, but not the end. Always extending the marker until the end of the line leads to undesirable highlighting in editors, for example:
Starting an "unused argument" diagnostic at
unused_argand extending It until EOL would also highlightused_arg, even though the diagnostic doesn't apply there.With this commit, the end position sent to the editor is automatically determined by looking for an identifier at the indicated start position. If there is no valid identifier at that position, the entire line is highlighted as before.
Note that in some cases, Pylint doesn't supply the start position of the token of interest, but instead marks the first non-whitespace character. I noticed this on import statements and except clauses, where the
import(orfrom) and theexceptkeywords then are what gets highlighted, instead of the module or variable in question. However, I think these are only minor issues in comparison to the original issue with the example above.