As did I for a number of years. But going from etch to lenny, the config for my custom kernel that I've used for quite some time did not result in a bootable kernel via grub 2. Even taking the /boot/config from the debian kernel and modifying it to my minimalistic needs resulted in an unbootable kernel. Even though I was using the same kernel and same config from etch that I had used for months. Note: not the kpkg-make or whatever the package variant of a custom install. Note: 2.6.28.10 because I've got one of those old video cards (laptop) that ATI's proprietary drivers stopped working on since their 9.3 driver. Note: I generally only enable the features that I use, even though there is hardware with drivers that I don't use.

Since upgrading to Debian Lenny I've only been able to run Distro kernels. Simply because I cannot seem to boot any custom kernel that I can come up with. And I've been running and configuring custom kernels for years. Not that it matters that much on my current very dated hardware where the stock kernel does most of what I need it to do. (except low latency for audio or compiled specifically for my CPU).

localmodconfig

I'll have to look into that.

As far as packages. In past versions I've tried upgrading from a minimal install to testing and add packages from there. And certain rather common packages were no longer available like Evolution, Open Office, and other things when I went that route. At which point I either had to start from square one (being on dialup at the time) or just do the next upgrade to sid to obtain access to what I considered basic packages back then. I opted for sid, since I couldn't even get the testing versions of things from packages.debian.org and other headaches. Of course sid had about a 10MB average of updates per day at that time. Which is an hour commitment PER DAY over dialup to stay current. And if some things were in flux like open office or kde, eternal bandwidth never current loop from hell.

Fortunately I'm now on quasi-broadband (2 megs) and updating just the list of packages isn't a half hour commitment or longer anymore.

Did you get any pointers as to what was stopping your custom kernel booting?.

Yep,that's what I use.
Then make menuconfig after,just to make sure the configuration is sane.

Can't say I've had too many problems with packages missing from Testing,but as I run a mixed Testing/Unstable system,I can grab the occasional package from Unstable when necessary.

Nothing definitive from what I recall. It was the same .config I've used on etch for quite a while. Not that I've tried lately to be able to definitively say what warning it does or doesn't give. But it's not the usual VFS panic type of failure. I'll have to try again to be sure, but something along the lines of not being able to read the root fs. Which is probably more of a grub2 issue than a kernel issue. Not that I'd know why it'd have the least bit of difficulty with ext3. But I'll have to actually try again in recent history to know those details.

I had something similar a while back.
Turned out the system was not creating an initrd,I know some people do without them,but as Debian do it that way so do I.
I posted my solution over at the 'official' Debian forums:

http://forums.debian.net/viewtopic.php?f=10&t=52079

Maybe this could help your problem.