Testing libsatsolver on Fedora

James Antill james at fedoraproject.org
Fri Jul 31 19:46:33 UTC 2009


On Fri, 2009-07-31 at 14:21 -0400, Bill McGonigle wrote:
> On 07/31/2009 01:12 AM, James Antill wrote:
> >  *sigh*, if you want to do some benchmarking of different package
> > managers available in Fedora (zypp makes the 4th, if apt is working
> > again) then feel free to actually do _a bunch of work_ comparing apples
> > to apples. You'll almost certainly be speaking privately with developers
> > from all of the tools, to make sure you aren't screwing it up. _Then_
> > post the results somewhere.
> 
> Raising the bar so high on public discussion of application performance
> is essentially an attempt to stifle it.

 You're right, I deeply apologize for asking people to _do the work_
necessary to produce real worthwhile performance numbers ... instead of
just posting the first thing that comes to mind, so we can all have a
giant flamewar full of "yum FTW!", "no, solv FTW, you suck!"

>   We don't make progress by
> commissioning scientific studies as soon as new ideas are put forth, we
> see first if they pass the smell test.

 We also don't make progress by posting "yum is 50x slower than solv for
update", and yet _predictably_ that is what this thread degenerated into
within hours of your post.
 I was probably fooling myself hoping to stop that before it started,
and likely clinically insane by writing:

http://illiterat.livejournal.com/7412.html

...to try and give a more detailed explanation of the problem, but then
maybe in the future I can just post a link to that and hit delete
thread.

> >  If, however, you want to just post "yum is slow" feel free to not do so
> > on f-d-l. Likewise with quick "benchmarks" like this (which amounts to
> > the same thing, IMO).
> 
> Somehow you forgot to quote where I said yum was doing more work and it
> downloaded twice as fast.  It sounds like you're trying to project
> unrelated anti-yum sentiment onto my simple report of positive progress
> on Michael's part.

 Then why post the numbers, if you know they aren't
comparable/worthwhile/etc?
 And, yeh, I'd have at least worded my reply differently (if not just
hit delete thread) if you were the first person to ever post weird
numbers and call it a yum vs. BLAH benchmark. But that works the other
way around too.


 And to make this clear:

 I'd be very happy if Michael would take over
https://bugzilla.redhat.com/447740 or otherwise help them complete. And
I would be even happier if someone wanted to do the significant amount
of work to provide some real benchmarks for package management (which in
the case of zypper vs. yum would kind of be blocked by 447740, IMO).


-- 
James Antill - james at fedoraproject.org
"I'd just like to see a realistic approach to updates via
 packages." -- Les Mikesell




More information about the devel mailing list