On Thu, Feb 03/00, kim@... wrote:

> From: <kim@...>
>
> On Thu, 3 Feb 2000 jhenders@... wrote:
>
> > Another interesting thing I haven't seen anyone comment on. Last
> > (emergency) patch they said they had to put in some changes half
> > finished when they put in the fixes for the network code, i.e. the necro
> > pet changes. Don't any of the programmers on the list find it odd that
> > they don't appear to have some kind of source control system that would
> > allow patching of the currently running code so they didn't have to go
> > live with partly finished changes?
>
> You're proposing they bring down the servers again to repatch.
> They may have decided that disruption was not worth the
> benefit of temporarily removing a nerf to one of the strongest
> classes in the game. OTOH, when the bard proc bug showed up
> again, they did an emergency patch two days later.

No, you misunderstood. They brought the servers down to patch for some
crash bug last weekend. The patch note when they came back up said
they'd also put in the partially completed necro changes as these were
part of the code base they'd fixed the crash bug in.

>
> Also, in large software projects, you tend to go with tested
> executables rather than pieces. The executable they tested
> had the necro pet nerf in it. Rolling back just that one
> piece of code may have caused unanticipated and untested
> problems in other sections of the code.

Tested? I'd be willing to bet the only part of that patch that was
tested was the part addressing the crash bug. Also, my experience with
large projects is usually that they use some sort of version control
system where merging in the fixes for one bug is easy to do by checking
out the last major revision. Patching the current code base, where there
are also many partial or untested changes is more a sign of no version
control system, IMO.

--
Artificial Intelligence stands no chance against Natural Stupidity.
GAT d- -p+(--) c++++ l++ u++ t- m--- W--- !v
b+++ e* s-/+ n-(?) h++ f+g+ w+++ y*