Re: Why does coroutine.create return a thread?

[Coda Highland <chighland@...>]

On Tue, Jun 21, 2016 at 1:48 PM, Gregg Reynolds <dev@mobileink.com> wrote:
>
> On Jun 21, 2016 3:40 PM, "Coda Highland" <chighland@gmail.com> wrote:
>>
>> On Tue, Jun 21, 2016 at 1:37 PM, Gregg Reynolds <dev@mobileink.com> wrote:
>> >
>> > On Jun 21, 2016 3:31 PM, "Coda Highland" <chighland@gmail.com> wrote:
>> >>
>> >> On Tue, Jun 21, 2016 at 1:27 PM, Gregg Reynolds <dev@mobileink.com>
>> >> wrote:
>> >> >
>> >> > On Jun 21, 2016 3:02 PM, "Coda Highland" <chighland@gmail.com> wrote:
>> >> >>
>> >> >> On Tue, Jun 21, 2016 at 12:45 PM, Gregg Reynolds <dev@mobileink.com>
>> >> >> wrote:
>> >> >> >
>> >> >> > On Jun 21, 2016 10:45 AM, "Roberto Ierusalimschy"
>> >> >> > <roberto@inf.puc-rio.br>
>> >> >> > wrote:
>> >> >> > ...
>> >> >> >> Right: couroutine is a specific use case of a Lua thread. As it
>> >> >> >> is
>> >> >> >> the
>> >> >> >> only use of threads inside Lua code, we can say that, looking
>> >> >> >> from
>> >> >> >> Lua,
>> >> >> >> coroutine = thread.
>> >> >> >
>> >> >> > Is it correct to say that a Lua thread is a fiber?  I.e. not an OS
>> >> >> > or
>> >> >> > pthread?
>> >> >> >
>> >> >> > - gregg
>> >> >>
>> >> >> Yes, that could be considered accurate, as fibers are a
>> >> >> cooperative-multitasking construct.
>> >> >>
>> >> >
>> >> > the amazingly cool thing is that we can have cooperative multitasking
>> >> > even
>> >> > with a single sequential thread of execution.  magic.  kinda makes
>> >> > you
>> >> > rethink your concept of execution,  thread, etc.
>> >>
>> >> The fun part is when you realize that before Windows 95, the entire
>> >> Windows operating system was cooperatively multitasked, and apps were
>> >> expected to yield quickly when awakened by the OS.
>> >>
>> > MS-DOS: the Original Embedded System.
>>
>> MS-DOS wasn't even cooperatively multitasked. It was TECHNICALLY
>> preemptively multitasked, in that triggering a TSR through an
>> interrupt routine could happen at effectively arbitrary times in
>> program flow, but there was no timesharing and the TSR would have to
>> clean itself up and return control to the original program on its own.
>>
> Oh man, you're taking me back!  The last time I heard "TSR" was about the
> last time I heard "HCF" (Halt and Catch Fire, for the mainframe-naifs among
> us).  But that means Windows was never merely cooperatively multitasked, no?
>
> I wonder, was there ever a purely cooperatively multitasked OS for PC-like
> systems?
>
> g

Windows didn't pull the TSR stunt and the only interrupt handler it
didn't return quickly from was ctrl+alt+del. You can effectively
consider Windows 3.x to be cooperatively multitasked and MS-DOS to be
non-multitasked because the exceptions are nitpicky details that are
effectively only there for the sake of supporting a broader set of
hardware.

/s/ Adam