20785, "bradcray", "Understand impacts of new readLine() interface on k-nucleotide benchmarks", "2022-10-06T19:42:19Z"
When we switched to the new readLine() interface in #19937, we took a performance hit to k-nucleotide versions that switched to the new interface as seen in Chapel Performance Graphs The submitted versions initially took a hit, but then this went away when we reverted the change to them in order to have them continue to reflect the live versions on the CLBG website.
These new versions also took a code size hit (arguably a productivity hit?) in that cases where we needed to use one readLine() call per loop before now require two.
For both these reasons, I've been hesitant to have us remove the deprecated forms, at least without an understanding of why the slips happened and an agreement that we're OK with them / prefer the new interfaces despite these impacts.