On Fri, Dec 16, 2011 at 8:20 AM, Meikel Brandmeyer <m...@kotka.de> wrote:
> Hi,
>
> Am 16.12.2011 um 12:17 schrieb Cedric Greevey:
>
>> You *already* can't change the type signature of an optimized function
>> without needing to recompile its callers.
>
> But once you settled on a type signature, you have the usual fast 
> benchmark,optimize,reload-cycle. Without interfaces you don't even have that. 
> So it is worse in a sense.

My usual testing case is to directly call the function being tested
from a repl expression, which is recompiled if it's resubmitted.

Further, the "baked-in optimizations" can maybe be made togglable. The
unoptimized but easily reloadable 1.2 version would be called while
you debugged it; you could switch on optimizations just for
benchmarking the speed and tuning that.

>> Hasn't posting this discussion here "advanced the discussion"? Or are
>> even verbal suggestions not considered by the developers unless they
>> come from someone with a CA? If the latter, I'd suggest that that's
>> rather silly, only presumably said suggestion would promptly fall on
>> deaf ears. :)
>
> Posting this discussion certainly “advanced the discussion.” But an email 
> thread only gets you so far. Details get lost deeply in the response tree or 
> are spread over several messages. At some point, the thread has to be 
> distilled into some summary of discussion, outlining the motivation why this 
> is needed, what the different implementation approaches are, their pros and 
> cons, the alternatives, etc.
>
> The developers are certainly not deaf for suggestions, but providing them a 
> summary is certainly in everyone's interest. I'd certainly prefer Rich and 
> Co. to work on Clojure rather than wading through a long email thread.

Yes, but someone else who was participating in this thread and has the
details fresh in his mind can surely do it with less hoop-jumping and
more knowledge of the procedures than I can. Especially as at this
point I don't even know what a "CDD" even is ... :)

-- 
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your 
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en

Reply via email to