From: Bruno H. <br...@cl...> - 2004-06-08 18:18:04
|
Sam wrote: > so, for now they are ignored, right? Yes. > maybe we should save them and add a test to PUTHASH - IIUC, PUTHASH is > not speed-critical, at least not to the degree that GETHASH is. Yes, that's an idea. Since TYPEP can be quite expensive, it should be possible to turn on and off, and I think off should be the default in compiled code. (Like type checks through THE are turned off in compiled code.) So, what concrete mechanism for turning this check on and off do you propose? > BTW, you did not document them in the impnotes either. Because I don't like to document no-ops. > >> (and why didn't you bump O(version)?!) > > > > Because it's easier for you to do 'rm -f *.fas' than for me to do a > > cvs commit. > > I am sorry, I don't quite understand this. > I don't see why bumping O(version) is any different from writing a > proper ChangeLog or NEWS entry or checking for GC-safety or not mixing > up formatting and substantial changes. If I spend a minute changing O(version) every day, at the end of the year that makes three wasted hours. The same effect can be achieved by bumping it just once, before the release. Your analogy with the NEWS file doesn't hold because the NEWS file is a cumulative effort that grows over time. Bruno |