[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: Re: Use byte-code directly?
- From: David Given <dg@...>
- Date: Fri, 17 Sep 2004 22:52:57 +0100
George Warner wrote:
Nope. All the changes are flattened into one revision so that when we diff
the production builds we only see what actually changed between those two
builds. All the original individual check in/out comment(s) are compressed
into a single note. Nothing is lost except anything that was changed again
before the final production build.
Apart from all the revisions between your production builds, if I'm
understanding you correctly; if one production build uses version 1.4 of a
foo.c, and the next is 1.8, then code that gets added in 1.5 and then removed
again in 1.6 is gone forever, right?
If so, I think you probably ought to look at CVS tags. They'll let you do what
I think you're trying to do without actually losing anything. Tags are
basically sets of (file, revision) tuples. For example, the build-1 tag might
contain version 1.4 of foo.c, and build-2 can contain version 1.8. You can
check out a complete tag with 'cvs co -r build-1 myproject' at any point, and
it'll just work, regardless of what's been checked in on top of it... and,
more to the point, you can diff against a tag. So you don't get the CVS noise
unless you actually need it; but it's still there, if you want it.
But this doesn't have a lot to do with Lua...
function <zebnig:2368> (11 instructions, 44 bytes at 0x11fc50)
1 param, 9 stacks, 0 upvalues, 5 locals, 2 constants, 0 functions
Sorry, way beyond me. I know it's a for loop calling dprint(<var>, ":",
<var>), but you'll have to find someone who understands more about the VM.
How do you feel about COME FROM?
Come again? ;-)
[insert interesting .sig here]