On Fri, Mar 01, 2013 at 11:53:19AM +0100, Ralf Corsepius wrote:
On 03/01/2013 11:20 AM, Michael Schwendt wrote:
>On Fri, 01 Mar 2013 00:46:52 +0100, Kevin Kofler wrote:
>>Matthew Miller wrote:
>>>Historical footnote: I believe it was initially added to help squeeze the
>>>boot portion of Anaconda onto floppy disks. If the things you list are
>>>really the only things using it, I think it's time to retire it
>>The package was actually imported by Enrico in 2005.
>>If what you say is correct, that would mean it was already retired and he
>I must admit that I've forgotten what review procedure had been used
>in 2005. Only have found this odd thread:
> RFE: dietlibc review
>Who remembers the details?
I just found this:
I also vaguely recall some controversial reviews, where Enrico
insisted on using dietlibc instead of glibc for "performance
I recalled this set of issues too from my previous time in fesco but I
didn't find the meeting logs with the information. I did find this meeting
where fesco voted to disallow static linking to dietlibc but deferred the
question of linking to dietlibc at all.
On that question, I would tend to agree with patrice's email that we've
moved towards certain core systems being too core to let people use an
alternative in Fedora (although alternatives may be provided). Examples are
kernel (no kmods) and C compiler (IIRC, there was discussion about building
with clang that resolved in at least a decision on the list to use gcc).
There is a line somewhere as to what is "core enough" to warrant this
treatment but I think it's reasonable to think that the libc implementation
falls on the same side as the C compiler.