On Fri, Mar 15, 2013 at 11:34 AM, Stephen Joyce <sjbupc AT gmail DOT com> wrote:
> On Fri, Mar 15, 2013 at 11:47 AM, Les Mikesell <lesmikesell AT gmail DOT com>
> wrote:
>>
>> But the experience here says that
>> changing complex code requires extensive testing before trusting it.
>> It isn't a matter of not imagining this use case - it is going beyond
>> and imagining if  something goes wrong, or if you had to take over as
>> a replacement sysadmin where someone had made one-off changes like
>> that and even if they work you won't be able to get any help
>> understanding or maintaining them.
>
>
> No one ever said that the changes wouldn't be tested. I'll probably run the
> new system for weeks, watching the logs carefully and making any necessary
> changes, before trusting it.

Looking at a few logfiles isn't exactly what I'd call 'extensive'
testing.  Just off the top of my head, I'd wonder if the code that
does the collision-chain fixups still works as intended - or how to
verify that.  Rsync-based stuff will probably work - or appear to -
even if you completely break the pooling because it will link against
the previous copy in the pc tree anyway, but other methods may be
different.

> On Fri, 15 Mar 2013, Adam Goryachev wrote:
>> custom
>> versions of software always lead to pain (in my experience), so unless
>> there is some major commercial advantage, then I steer clear.
>
> Interesting point of view. I don't agree. I always considered the ability to
> make changes to the software to make it do what *I* want it to (rather than
> succumbing to what the original developer thinks I want it to do) as one of
> the hallmarks of opensource software. If you want to use software as-is,
> that can be beneficial too:
> http://news.slashdot.org/story/13/03/13/2052226/why-freeloaders-are-essential-to-foss-project-success

Regardless of what the youngsters on slashdot say, what I consider
most important about opensource software is how widely tested it is,
and the fact that by browsing the changelogs you can see the thousands
of subtle bugs that large projects have already fixed so they won't
affect you.  If you've been using free code for decades you will
realize that it is the bug reports and fixes that make it usable as
much as the original developer's code.

Sure, everyone is free to take their own chances with special-case
changes, but when you balance it against the risks, I don't see a big
win here unless your change goes into the mainstream code, which
doesn't seem likely at this point in backuppc's state.   Virtual
machines already provide a well-tested and understood solution to the
problem of sharing some resources with some level of isolation - with
some very usable free implementations.

-- 
   Les Mikesell
      lesmikesell AT gmail DOT com

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar
_______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List:    https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:    http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/