Hey,
>Oh gosh, I fear I've made it sound as though there's some exclusive
>membership requirement for this club or something, sorry. We're not
>trying to set up barriers and gates here; the only reason I'm
>recommending this course is because I think it's the fastest way to
>get these changes incorporated safely.
>
>
>
No I don't think that at all. I've been on the list long enough to
respect the talent. I'm just trying to be respectful. If my changes
weren't so large I wouldn't sound so cautious. And don't worry, you
won't scare me away. I'm stoked. When I'm stoked I code. When I code,
I smile:-) I don't want to freak you guys out but I have 6000+ LOC more
hiding in the wings. Don't worry they can be kept separate very
easily:-) I talk about it a little on the SQL page. I plan to first
clean and shrink it, then I'll add back in. That was the idea behind
the indirection. To allow scary stuff to be written and tested with all
the great stuff you guys have without toppling the most data critical
portion of the code. The other key ingredient is that it must be able
to be working side by side with the baseline in the same executable. In
order to preserve the FS API, I want to be able to run the same
"blackbox" test cases side by side against separate repo's using the
same binary. I choose the name baseline carefully. If an experimental
FS impl can't match the baseline it is not an FS impl by definition. If
testing conflicts occur that point to a problem with the baseline then
it must be fixed before any other versions change/improve the behaviour.
This is the way I'm thinking anyway. The baseline has to remain to
keep the FS stable.
Feel free to abuse as needed. Please limit abuse for the next two weeks
though. I'm into this stuff so much I can't resist the temptation to
respond:-)
gat
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Nov 22 20:26:35 2002