[Yum] yum 0.8.6 testing

[Troy Dawson]

seth vidal wrote:
> On Thu, 2002-06-13 at 15:06, Troy Dawson wrote:
> 
>>Here's to the fastest coder at Duke :)
>>
>>Anyway, I finolly tested the debug levels.  I think it's almost right, and my 
>>suggestions here are totally up for debate.  (Because at the moment, the cron 
>>jobs do what I want)
>>
>>When you do an install or update, and something get's installed or updated, 
>>then it works correctly, both the d 0, and the d 1 levels.  It also works 
>>correctly if you are doing an update, but don't give it any name.
>>
>>But, when you do an install or update, giving it a package name, then it gives 
>>back a reply for both -d 0 and -d 1 levels.  In my opinion, the -d 0 debug 
>>level shouldn't give back any reply for this, because this isn't really a 
>>fatal error.
>>
>># yum -d 0 -y install nedit
>># yum -d 0 -y install nedit
>>nedit is installed and is the latest version.
>># rpm -e nedit
>># yum -d 1 -y install nedit
>>Installed:  nedit-i386
>># yum -d 1 -y install nedit
>>nedit is installed and is the latest version.
>>#
>># yum -d 1 update
>># yum -d 1 update nedit
>>nedit is installed and the latest version.
>># yum -d 0 update
>># yum -d 0 update nedit
>>nedit is installed and the latest version.
>>#
> 
> 
> 
> So something that allowed a gradation of error conditions?
> If you grep for errorlog in the source code you can see all the the
> things that generate error conditions (sent to stderr)
> 
> I was having trouble judging which ones should result in an error
> statement.
> 
> it wouldn't be hard to add a -e # option - to do the same as debuglevel.
> 
> A little explanation - I had used Michael's logging class so there were
> 2 types of output - stdout and stderr.
> 
> errors go to stderr, notices,informational messages to stdout. Obvious.
> I have all errors with the same error level. I don't think I want
> debuglevel and errorlevel linked. I like the flexibility of knowing
> anything that is an error but I don't want to see informational
> messages.
> 
> so would a -e # work well enough for you?
> 
> -sv
> 
That sounds good to me if it isn't too much trouble.

Troy

-- 
__________________________________________________
Troy Dawson  dawson at fnal.gov  (630)840-6468
Fermilab  ComputingDivision/OSS  CSI Group
__________________________________________________