New rpm version about to hit rawhide

[Andy Green]

Paul Nasrat wrote:
> On Tue, 2007-06-26 at 09:11 +0100, Andy Green wrote:
>> Panu Matilainen wrote:
>>
>>> Expect memory use go up, dramatically. That's the price to pay for
>>> finally "fixing" the multilib %doc etc file removal bug by nuking
>>> the nasty skipDir hack that caused it to begin with. Whether we can
>>> actually live with the ballooned memory usage, especially in the
>>> installer, remains to be seen...
>> Of the two issues, surely memory use is the worst?
> 
> Many people seem to disagree.

I guess those 'many' people aren't installing and updating on low end
machines from a couple of years ago.

>>   I installed F7 on my
>> in-laws' old laptop with 256MB last week, it did install okay because I
>> took the precaution of turning everything except the most basic stuff
>> off (including X), then yumming it in afterwards.  Previously I upgraded
>> FC5 on a 384MB box and it took most of the day.  Are the "dramatic"
>> increases only seen in the multilib case itself?  That wouldn't be so
>> bad since x86_64 boxes will have lots of memory...
> 
> No the issue is not purely on multilib occurs with multiple identical
> basenames - eg multiple kernels, COPYING etc.

What I meant is, if the increases in memory footprint are only seen when
libraries from multiple arches are actually being dealt with, it
wouldn't be a regression for the older x86 boxes with relatively low
memory, nor such a problem for x86_64 boxes which typically have lots of
memory.

But if the price in memory for solving a multilib issue has to be paid
for even when a single arch is all there is, that would be more painful
since already struggling low memory single-arch boxes have to pay it for
no benefit to them.

-Andy