Show quoted text> Would you agree with me that we are now long past the days of limited
NO!
Show quoted text> memory and limited CPU speeds and that the tiny savings of NOT cacheing
> lexicals in symbol tables is, sort of, like behind us now? I mean, I
> cannot imagine that this simple operation could be a huge overhead hit
> these days, not when even the cheapest processors can perform very
> complex operations like virtualization! So, this seems to be a vestige
> of the thinking that went into making Perl very efficient... circa early
> 2000s at the longest, perhaps.
A good read on the topic is Small memory software by James Noble and
Charles Weir. It's a good read in that it simply deals with interesting
algorithms and recipes without jumping to conclusions about where we
stand.
Show quoted text> I might propose to the Perl developers that they add a facility for
> lexical cacheing either to a separate set of symbol tables so as not to
> break existing code, or provide a separate facility entirely. Maybe
> these could be command line options that one might enable for deep
> debugging?
I would suggest to approach any problem from the other end: let people
know what you want to achieve, and ask them what tools we they
recommend, and if there are no satisficing tools, how they would suggest
going forward. Especially in the language that has such a huge
repository of software it is often better to first ask for directions
that have already been explored. Did you even state what kind of
debugging you're after?
Show quoted text> Again, I know you are not part of that realm of the Perl software
> projects, but being you support a very related set of software, I
> thought you would be a good choice of someone to kick this idea back and
> forth with before pestering the folks in the Perl interpreter support
> area. At any rate, I value your opinion.
Thanks for your trust, sorry if I disappoint you with this answer:)
--
andreas
