Re: anyone notice speed of 2.17.95 on Windows ?

[David Kastrup]

Devon Schudy <address@hidden> writes:

> Keith OHara wrote:
>> the pure-estimate and unpure-final versions of a function
> [...]
>> The word 'pure' might have too much a connotation as 'good'.  Maybe we
>> should rename 'pure' ->
>> 'shitty_hack_estimate_because_I_am_unable_to_order_layout_decisions_better_please_forgive_me'
>
> Oh, so *that's* what all the 'pure' stuff is about! I was wondering.
> The CG says it's just functional purity, but that obviously isn't the
> whole story. Renaming them to something like 'estimate' would be less
> confusing. ('Raw' is opaque to me, but 'estimate' is clear. Or are
> they really ideals instead of estimates?)
>
> Is everything with 'pure' in its name for estimates, or are some of
> them just functions that happen to be pure, or that need to be pure
> for some other reason?
>
> Unpure-pure-container is also confusing. There's an explanatory
> comment at the top of unpure-pure-container.cc, but it's unfinished:
> “Used for rerouting a function of (grob start end) to one of (grob)”.

That's just for the special use case where only a single callback is
placed in the container.  Then a proxy function is created that throws
the extra arguments away.

> Perhaps it should be named after its purpose rather than its
> implementation: something like 'estimator', 'with-estimator',
> 'estimable', 'two-phase-layout-function'?

There were a few proposals already, most of them I like better than
these.

-- 
David Kastrup