Front page | perl.perl5.porters |
Postings from July 2009
On Wed, Jul 08, 2009 at 12:46:56AM -0700, Joshua ben Jore wrote: > On Wed, Jul 8, 2009 at 12:37 AM, Paul Johnson<paul@pjcj.net> wrote: > > Yes, a test is failing because the optree has changed somewhat and the > > Deparse code I nicked doesn't know about that. I assume that Deparse > > itself has been updated so I just need to nick the updated code. > > (There's that word again - "just".) > > I eat a live kitten every time the optree changes. We demand rigidly defined areas of doubt and uncertainty. If there is no freedom to change anything, then nothing can change. Which means that no bugs can ever be fixed, and no optimisations made. Not changing the optree in any fashion is equivalent to asking that a Java compiler continues to emit the same sequence of bytecode ops for the same input Java syntax. I appreciate that it causes local pain when it changes, but I think that the net global pain is less. Nicholas ClarkThread Previous | Thread Next