Florian Festi wrote:
> We used the following test cases:
> 
>  * normal system (~1200 pkgs)
>   * empty update (no updates)
>   * small update (3-5 updates)
>   * big update (150-200 updates)
>   * dist upgrade to next F(C) release
> * install into chroot
>  * normal install (1200 pkgs) - use groups, not a pkglist
>  * full install (7500 pkgs)
> * erase glibc
> 
> These cover most of the runtime behavior.
> 
>> if for no other reason that to prove that we've done _something_ in all
>> these changes. :)
> 
> I doubt that this will stay the only reason. Improving is difficult
> without measurement. Without testing the big cases quadratic, cubic or
> worse runtime behavior stays undiscovered.

Agreed. So does anyone want to script up the above tests in a way that
they can be run repeatably by unprivledged users in an automated
fashion? Then they can be shipped with yum, and we could again consider
the idea of automated nightly performance runs.

-James

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Yum-devel mailing list
[email protected]
https://lists.dulug.duke.edu/mailman/listinfo/yum-devel

Reply via email to