On my system, "yes -- '-nan' | head -116903 | sort -g > /dev/null" completes in under a second. while "yes -- '-nan' | head -116904 | sort -g > /dev/null" appears to hang. The same behavior is observed for "+nan". For "nan", the border is between 120150 and 120151. An older version of sort, 8.5, appears to work fine.
Aaron Denney wrote: Thanks for the report. Technically, this is not due to a bug in coreutils, but I've patched sort.c, at least temporarily, to work around the problem: http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/23080
Aaron Denney wrote: Thanks for the report. Technically, this is not due to a bug in coreutils, but I've patched sort.c, at least temporarily, to work around the problem: http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/23080