17/06/2026 19:44, Matthew Gee: > This patch fixes a bug where review-patch.py would detect and report an > error or warning only based off of the occurrence of the headers of the > error and warning sections. This led to consistent false positives as > often AI reviewers will create the header but put "none" or similar > filler text within the following body. > > This patch updates the code in order to check if the AI review has a > body with error or warnings to fix and not just filler text. This is > done by keeping track of whether the for loop parser is within an error > or warning section; analyzing the first non-whitespace line within the > section. If the first non-whitespace line matches known filler then the > section can be ignored. It has been observed that if the AI includes > filler then there is no actual concern. > > These changes were tested against 10+ markdown AI review outputs with > several variations in formatting and filler text. The changes caught > error or warning sections with actual concerns and successfully ignored > sections containing only filler. > > Signed-off-by: Matthew Gee <[email protected]>
Applied, thanks.
