develooper Front page | perl.perl5.porters | Postings from August 2021

Re: PSC #033 2021-08-13

Thread Previous | Thread Next
Philip R Brenan
August 15, 2021 23:42
Re: PSC #033 2021-08-13
Message ID:
The Intel Emulator:
gives an exact count of the instructions executed by a program.  Please
consider using the number of instructions executed (amongst many other
useful statistics)  to see whether a proposed change is going to increase
or decrease the number of instructions executed and thus have a possibly
negative or positive effect on execution speed?

Please tell me how you intend to perform the action:  *Communicating beyond
p5p  *so that I know where to look for such information?

On Mon, Aug 16, 2021 at 12:27 AM Neil Bowers <> wrote:

> PSC #033 2021-08-13
> Present: Paul, Rik, Neil
> *Source encoding stricture*
> We had a follow-up discussion on the proposal that came out of last week's
> meeting, for a source encoding stricture[1]. One issue raised on p5p was
> the question of having a file with two different encodings in it. This is
> independent of the proposal – if you `use utf8` now, you still have this
> problem. And is this a real life problem? We don't think so.
> The other topic was the fact that if you have non-ASCII UTF-8 in your
> source and the pod, then you need `use utf8` and `=encoding utf8`. Naively
> it would be nice if you could get away with just `use utf8` at the top of
> the file, but pod parsers aren't (and shouldn't be) expected to parse the
> non-pod parts of your source, and the utf8 pragma could be used by some
> module you're using, and so wouldn't be seen by the pod parser anyway.
> The proposal doesn't fix everything, but it's a step in the right
> direction. We'll let the discussion another week or so, then make a
> decision on whether to go to a formal RFC.
> [1]
> *Communicating beyond p5p*
> On p5p we're discussing things that are going to affect all Perl
> programmers, but often they're not aware of what's coming until it's
> released, or about to be. For most discussion points that's ok. But we
> agreed that for some topics the PSC should let the broader communities know
> of changes being considered, to give people a chance to comment, and at
> least set expectations.
> An example of this is the change we're considering with respect to
> taint[2]. Neil will do a short blog post on this.
> [2]
> *PV vs IV/NV discussion and PR*
> Following the discussion on improving how Perl decides whether to render a
> scalar as a string or number[3], we're keen to see Nick's PR[4] merged, but
> before we can do that the last concern is performance. We don't have tools,
> or even an agreed process on how to benchmark changes. This applies to the
> proposal for dropping taint – we believe this will bring a good performance
> boost, but how confident are we, and what kind of things will see the
> benefit?
> Paul will talk to Nick to come up with something for this specific case,
> which we can then hopefully look to generalise.
> [3]
> [4]
> *Lord of the Quirks*
> We continued our review of the quirks document, but didn't get through too
> many, as the previous discussions took up most of the hour that was
> available today.
> One of them was use of ' as package name separator. Rik has already kicked
> off that thread on p5p.
> Neil

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About