in reply to Performance penalty of using qr//
FTR: I encountered (and probably described/reported it here, but I don't remember for sure), a similar disappointment many years ago.
My recollection of it is that a variable built using qr// was always reinterpreted at the point of interpolation into another qr// or a larger regex.
My conclusion was that like the /o modifier on regex, it was either never implemented, or was removed by an early bug fix.
See also 971076, but I'm almost sure there should be an earlier thread.
|
|---|