file locking primitive
Tim Rowledge
tim at sumeru.stanford.edu
Sun Jan 4 00:16:50 UTC 2004
"David T. Lewis" <lewis at mail.msen.com> wrote:
> I think that the approach you describe here is pretty much exactly
> what I did for OSPP, right? I used #shouldBeTranslated to figure out
> what subclass to generate, allowed significant differences between
> the Windows and Unix plugins, and relied on OSProcessAccessor and
> its platform-specific subclasses to sort things out as much as possible
> on the image side.
Pretty much. I should have remembered and used it as an example!
> The reason I ask is that the design evolved into
> this approach totally haphazardly, and I'd be interested to know if
> I accidentally got it right.
I think it has more to do with being a decent software engineer than
accident..
> It does seem to play nicely with VMM too,
> although we would have to give it a little more thought if we wanted
> to be able to cross-compile the plugins for Windows from a Unix Squeak
> and vice versa (which should be entirely possible, although I've never
> actually tried it).
True; I suppose the VMM stuff would have to provide some access to the
chosen target platform name beyond what it currently does. Another
issue with cross-generation is our old fiend(sic) lineends. Yet another
reason to write some filestream class with enough brains to open the
windwos before puking in the car.
tim
--
Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
"How many Pak Protectors does it take to change a lightbulb?" "Only one, but the lightbulb has to smell right."
More information about the Squeak-dev
mailing list
|