On 20-Nov-05, at 9:43 PM, Nigel Pearson wrote:
but I don't think it guarantees that the resultant code
will interoperate. r1234 of mythtv might not talk to r1234
of mythplugins, and the situation might not change until r2345.
Since svn revision numbers are repository-wide it does guarantee
consistency from a source control point of view. Sure, a clueless or
malicious developer could commit intentional breakage but I didn't
think we had any of those... unless you'd like to name names ;)
Point is mixing revisions is wrong. You can argue that most of the time
it's only a little wrong but it's still wrong. Closing that obvious
hole makes things a little more robust and that's good. Same deal for
cleaning nulls out of System arguments. There's no technical reason to
leave potential problems in there and they always end up biting you
when you're in a hurry and you don't have time to deal with them and
your wife/SO is yelling at you to finish with the computer so lets code
defensively and live happily ever after.
because it regenerates some headers, which will force a big
re-compile of most of the source.
A workaround would be to add a "rebuild" option,
which skips the reconfigure (and maybe the Makefile deletion).
OK. I figured there were too many options already but if that's the
best approach... I'll revisit in-place builds when I get some time in
December.
I'm having trouble configuring with the single disk image but that's
probably because I've been copying the built dependancies from one
image to another and something must be out of place. Tomorrow I'll do a
full run from scratch and when that succeeds I'll post an updated
packager script.
- George
_______________________________________________
mythtv-dev mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
