On 09/06/2013 02:42 PM, Nicholas Clark (via RT) wrote: > 1) nothing will break if we removed _qsort. ... on CPAN. > but it's not exactly high maintenence code. It's certainly not, nut pp_sort.c as a whole is moderately complex. Reducing the amount of code there can only be good. > So, should we put _qsort through a deprecation cycle, and drop it in v5.23? > > > I'm not actually sure if it's worth it. Do the small gains outweigh the small > churn? +1 to see it go. I'm trying not to grandly suggest that somebody (not me, likely not Nicholas) should try implementing timsort for us now instead of the current merge-alike. :) --SteffenThread Previous | Thread Next