develooper Front page | perl.perl5.porters | Postings from January 2017

Re: Should we consider locked hashes a failed experiment?

Thread Previous | Thread Next
Leon Timmermans
January 28, 2017 15:44
Re: Should we consider locked hashes a failed experiment?
Message ID:
On Sat, Jan 28, 2017 at 3:31 PM, demerphq <> wrote:

> My experience with locked hashes is that as currently implemented they
> are nearly useless. A simple read of a non-existent key becomes a
> fatal exception, which makes them nearly unusable in conventional
> code. At work we have tried to use them for a few purposes, and
> generally it is has been a failure.  They impose a run time penalty on
> all hash access even though they are barely used.
> As far as I can tell locked hashes were invented as a way to work
> around our lack of core support for true objects/structs. (Where
> accessing a non-existent member *should* throw an exception, but
> preferably a compile time one.) However with perls standard object
> model and usage patterns I have never really seen anyone use them this
> way.
> Steffen has started working on a patch series to make hashes
> pluggable. (Big big fricking A to him on that!) This is only going to
> be more complex if we support locking (does every hash implementation
> have to support locking? Does the hash api have to support
> "clear_placeholders"?
> I propose that we get rid of locked hashes. I have written tons of
> code to help support them, and I consider them a total failed
> experiment.
> Also, use of clear_placeholders and related functionality is one of
> the "hidden apis" which we have in Internals that really has no
> business being exposed to end users.
> Once we have a pluggable API for hashes, we can implement lockable
> hashes using them and if people really want them we can provide them.
> What do people think? Should every hash access and store pay a penalty
> for locked hashes when they are almost unused?
> Yves

As far as I know they're only used in the fields implementation, and were
introduced when pseudohashes were abolished. I believe readonly hashes are
generally far more useful, and that fields can be implemented quite easily
on top of a pluggable hash API.


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