[DGD] FLS

[Raymond Jennings]

Blain, I'm just curious.  What was your intended use case for FLS?

I also just remembered you could just fake it by using normal TLS that has
a sub-array that is indexed by frame.


On Thu, Oct 5, 2017 at 12:25 AM, Raymond Jennings <shentino at gmail.com>
wrote:

> At any rate you should worry about correctness before you worry about
> efficiency.
>
> Get it working right, then worry about optimizing
>
> On Wed, Oct 4, 2017 at 11:36 PM, Felix A. Croes <felix at dworkin.nl> wrote:
>
>> Blain <blain20 at gmail.com> wrote:
>>
>> > I implemented the this_user idea in the kernellib's TLS implementation,
>> but
>> > there was no easy way to tell if the setting corresponding to a given
>> frame
>> > is expired.  So thus the FLS idea was born.
>> >
>> > If DGD could supply a null element before the first arg, all else could
>> be
>> > handled by the lib.  The current klib's implementation of TLS could
>> also be
>> > made to use this instead of having to create wrappers for entry-level
>> > functions in driver.c to basically do the same thing (except with only
>> > frame 0).
>>
>> This_player() functionality has been implemented entirely in LPC before,
>> without using FLS or even TLS, and that implementation was even more
>> efficient than what you described since it doesn't involve looping
>> over frames.  I see no reason to hack DGD for it.
>>
>> Regards,
>> Felix Croes
>> ____________________________________________
>> https://mail.dworkin.nl/mailman/listinfo/dgd
>>
>
>