Two questions concerning bootstrap ...

[Imago]

Hi
while playing around with the bootstrap script for almost the whole day i discovered two things which seems quite wrong for me.
But before filing another bug report I just want to verify that it isnt just a thinking error of me

a) concerning the --resume option
As far as i understand it should do the following:
running bootscript.sh with --resume option build the binary .tbz2 package for each emerged component during bootstrap. So when you have to interrupt the script or it interrupts itself, you can restart it using the --resume option and it will install the binary package instead of compiling everything again from scratch. So you can almost start from the point you have aborted.
Is this so far correct?
if yes this has a side effect which renders bootstrap quite useless.
Because as ive seen the script bilds first a toolchain(glibc/gcc/etc) and then uses this toolchain to compile the actual base system. Again is this so far correct?
If yes, we're compiling several packages twice. But when using the resume option things dont get compiled a second time instead the binary package from the first compilage is installed. So you actually only have done a 1/2 bootstrap

b) concerning gcc profile switching during bootstraping
As far is I have looked into it, when emerging new gcc package during bootstrap the ebuild will automatically call gcc-config and sets it to the new profile. At least it is done so for gcc >=3.4 which uses the routines from the toolchain.eclass. So far everything is correct, but i found nowhere a point where a 'source /etc/profile' or something similiar is executed.
So, although the profile is correctly set to the new gcc version, the environment variables still point to the old one.
I verified that by putting some 'gcc -dumpversion' into the bootstrap script which shows that although emerging gcc-3.4.3 during bootstrap everything is still compiled using the gcc-3.3.4 from the livecd.
Or is it meant to be that way?

CU
Imago