On Tue, Sep 11, 2012 at 10:00 PM, zxq9 <[log in to unmask]> wrote:
> On 09/12/2012 04:19 AM, D Brandherm wrote:
>>>
>>> Looks like we've found the source of the problem, I'd take these logs to
>>> the retrospect folks.
>>>
>>> Pat
>>
>>
>> Sounds like a plan. Thanks again to everybody for your help with this.
>>
>> Dirk
>
>
> Wow... 44G worth of log spamming over a password request. So that's what a
> modern resource exhaustion crash looks like. Orders of magnitude huger than
> the same sort of error 20 years ago!
>
> I'm glad to hear you got everything straightened out.

Heh. I went through similar nuttiness with SuSE's "YaST" tools, which
required manual interactions for NVidia driver handling, Java
installers from Sun, and other software builders and installers that
insisted on whinging to the installer. Even ran into one in a CPAN
tool the other day, had to tell the .spec file to run "yes | perl
MakeFile.PL" to get it to build the testable RPM's

> The Fedora/RHEL rule is that no package installation process may ever
> require user interaction as part of the installation. If interaction is

Yeah, and that's part of why Sun refused to properly RPM bundle their
Java toolkits: they wanted the installer actively agreeing to the Sun
license. That's also why I'm so glad Oracle has thrown that out and
gone to OpenJDK's much more workable licenses, with honest-to-god
SRPM's and sanely named packages.