Quote:
Originally posted by jaguar
stil lbest to have hte latest build, or milestone
|
Not always. Trust me. :-)
After more than thirty years working in software development, seven of which was direct involvement as a support engineer, I can say with confidence that immediately upgrading to every new thing that comes down the pipe is *not* always the best strategy to maximize stability and minimize effort. There's an art to pickling out the sweet spot as to how much/often to upgrade.
Naturally, when I was dealing with customer issues, I often would urge them to get uplevel if they were backlevel...especially seriously backlevel. But in most cases we were dealing with issues whose cause was not clear, and we didn't have anywhere *near* the userbase Mozilla has. And furthermore if we managed to establish that there was actually a defect in the product, I would still have to get it to reproduce in *current* code in order to get the developer to look at the issue.
This is a different situation. I *like* my Mozilla in general. I have good workarounds for the bugs that have bit me so far. And the status of those bugs is *immediately* knowable on Bugzilla. If I pull down a new milestone, it takes time and effort on my part, and exposes me to a random collection of new defects that may be in the new code, which I'll then have to work out resolutions for. If I knew that 0.9.7 would have goodness in it that would make it worth my while, I'd pursue it.
My next big put-new-code-in-the-box project will be RedHat 7.2...at which point even though I know it fixes major issues in the crappy code RH shipped to manage dial-up ppp, I'll still fret about whether my nVidia display driver (which is actually a binary in RPM drag) and Equinox multport serial card driver will work correctly with the new kernel. There's no particular *reason* to think they won't, of course.