Lee Naish wrote:
> The "J" command seems to have a complexity bug when given a large numeric
> argument, eg 100000J
...
> ie, doubling the number of lines increased the runtime by a factor of 4
> exactly - seems like there is a O(N^2) component to the algorithm.
The "=" command also has O(n^2) complexity:
Here are some timing using the following test case...
$ for lines in 1000 2000 4000 8000 16000 32000 ; do
echo "=== $lines"
yes | head -$lines > yes-$lines.txt
/usr/bin/time vim -u NONE -c 'normal =G' -c 'q!' yes-$lines.txt
done
#lines timing
1000 0.25user 0.01system 0:00.33elapsed 79%CPU
2000 0.78user 0.04system 0:00.86elapsed 96%CPU
4000 2.91user 0.06system 0:03.30elapsed 90%CPU
8000 11.26user 0.24system 0:12.33elapsed 93%CPU
16000 46.47user 0.86system 0:47.83elapsed 98%CPU
32000 182.99user 3.76system 3:07.84elapsed 99%CPU
I.e. indenting 2*n lines is ~ 4x longer than indenting n lines.
-- Dominique
--~--~---------~--~----~------------~-------~--~----~
You received this message from the "vim_dev" maillist.
For more information, visit http://www.vim.org/maillist.php
-~----------~----~----~----~------~----~------~--~---
