Re: Glibc 2.2 and potato

[Michael Meding <Michael@Meding.net>]

Hi Colin,


> [Hmm, I think you sent this to both me and the list *in different
> messages*; please don't ...]

after I wrote the message I've seen that netscape replied just to you
instead of the list. Don't know what is wrong here, I get this when
replying to any kind of debian list. Hm...
that's why I copied and pasted and then sent to the list. Of course you
are right here.

> Buh, I'm being stupid - I was just arguing yesterday about why binary
> packages are OK and compiling from source isn't necessarily the One True
> Way to get software, and the wrong triggers fired in my brain. :)
> 
> It should give you the same results, yes, assuming you have the right
> build-depends. For using different kernel headers, you should be able to
> work out where they're expected to go by reading debian/rules etc. in
> the source archive. I'd be careful about applying aggressive
> optimizations; a lot of the higher ones can be a little risky. But it
> may be worth a shot if you believe libc6 to be a major bottleneck in
> your system. (Personally, I think disk I/O, memory caching, and the like
> are much more serious bottlenecks than the C library, but you can always
> try it.)
Thanks for the in depth thinking. Still I am thinking that optimizing
there should result in an improved speed.

Anyway, there should be something coming out of compiling glibc against
pre 2.4headers. What about the compatibility between 2.13 and 2.2 .
Should that work out (binary wise?) ? And what could one expect by
compiling against headers of 2.4 ?

TIA

Greetings

Michael