RE: toolchain requirements submission

["Simon Poole" <simon dot armlinux at themalago dot net>]

> -----Original Message-----
>
> <slightly disappointed rant/>
> 

I think you should post an arbitrary proposed spec for the toolchain and see
how people respond.  If it has problems, people will correct it, and if
no-one responds then you can take that as a blessing to proceed...
  
> Mulitlibbing has my preference, precisely because of this reason.
> 

The way I see it, we need to do things in the following order:
1. Decide which versions we are going to put in the toolchain
2. Produce mutiple toolchains (hard/soft, big/little) with those versions.
This step should be pretty trivial.
3. Then begin the more difficult step of getting crosstool to do
multilibbing or whatever.

The sooner we get step 2 completed, the sooner our standardised toolchains
can start being tested/used/patched by end-users.
Opinions?

Here's the toolchain spec I currently use for my Cirrus EP9312-based
hardware:
* gcc-3.3.2 (c,c++)
* binutils-2.14.90.0.8
* glibc-2.2.5
* --target=arm9tdmi --with-cpu=arm9tdmi
* little-endian, soft-float
I tried the same combination with glibc-2.3.2, but that gave me problems
with busybox (init would always segfault), so I stuck with 2.2.5 for the
time being.  I haven't checked whether a newer crosstool/gcc/binutils would
solve that particular problem.

Simon.
--
Simon Poole
http://www.appliancestudio.com/


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com