On 04/07/2011 22:45, Simon Hengel wrote:

Hi all, I realized that ghci shows non-linear time complexity for a sequence of :module and :reload commands, say we do multiple times:

:module *Foo :reload

Steps to reproduce:

time for i in {1..10}; do echo -e ':m Prelude\n:r'; done |ghci time for i in {1..99}; do echo -e ':m Prelude\n:r'; done |ghci

On my system the first run uses 1 second of CPU time, the second 39 seconds. Is this considered a bug?

Could you create a ticket please? That looks like an interesting result and we should investigate. Cheers, Simon