Problem
.github/benchmark-diff.sh claims to process CSV files but removes every double quote from each field and splits only on commas. A quoted field containing a comma is therefore parsed as multiple columns.
Steps to reproduce
Create a baseline or current benchmark row whose first field is quoted and contains a comma, for example:
Run benchmark-diff.sh baseline.csv current.csv 1.
Actual result
The first field is split at the comma, so the benchmark is assigned the wrong identifier and the comparison table is corrupted.
Expected result
Parse quoted CSV fields according to CSV rules, preserving commas inside quoted benchmark names.
Problem
.github/benchmark-diff.shclaims to process CSV files but removes every double quote from each field and splits only on commas. A quoted field containing a comma is therefore parsed as multiple columns.Steps to reproduce
Create a baseline or current benchmark row whose first field is quoted and contains a comma, for example:
Run
benchmark-diff.sh baseline.csv current.csv 1.Actual result
The first field is split at the comma, so the benchmark is assigned the wrong identifier and the comparison table is corrupted.
Expected result
Parse quoted CSV fields according to CSV rules, preserving commas inside quoted benchmark names.