[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

RE: Debian 2.0 Mac install feedback


>   I just completed the base installation process of beta 2.0 on my SE/30.
>I had no serious problems with the install process, of concern to me was
>that after configuring modules, dinstall told me that the base system had
>been configured and suggested that I configure the network.  Other issues

That's a problem that has been reported before. I admit I never bothered to
look at dinstall to find out what fools the process into thinking base was 
already configured ...

>I'd like to see addressed:
> * The dialog boxes and message files are expecting an 80 column console,
>but some Macs only have 64 columns.

I'd hope that someone debugs the 6x11 console code some day, which would give
you 80 columns. Randy Thelen, the author of that driver, seems to be terminally
overworked ATM... 

> * At a time when the only Linux native partition was already initialized,
>dinstall suggested Initialize a Linux Partition with a higher preference
>than Mount a Previously Initialized Partition.

The install and the kernel has no way of knowing if there's a partition 
already initialized AFAIK. Even if we keep a list of partitions that were 
initialized in that session, the user still might have initialized a partition 
in a previous session. We'd need to test mount all of the potential partitions
to figure out which has a valid filesystem.

> * The search for locations of install disk image files should be pruned at
>directories whose names begin with a dot (such as ".resource") or otherwise
>managed so as to skip the special HFS directories.

HFS brokenness. Either someone writes a AppleSingle option for the HFS code
in the kernel (not my top priority, neither asun's), or someone explains how I
can compile and build a patched dpkg. 

> * If kernel modules are not supported on the Mac, and none were installed,
>there's no need to configure them.

In the ideal world, there's a modularized Mac kernel for Debian. In the real
world, there's not even complete support for all SCSI chips on Macs, and the 
only existing SCSI driver seems to break on mke2fs for at least some Macs. 
I'd appreciate help in both the SCSI driver and the 'modularized kernel' issue

> * The system complained about hwclock errors when booting.

Kernel problem; the routines to read and write the RTC haven't been written 
(some Macs access the RTC via ADB, there should be pmac code to glance the 
required commands from. Others have the RTC attached elsewhere. No clue what
is required in those cases). If someone can point out what the code path 
on the hwclock ioctl is, I can add a dummy RTC handler and the write is 
silently ignored, reads can be filled from the scheduler clock.

>   I guess that's it for installer-related issues.  Like I said, no serious
>problems, and things worked as expected.  Great job!

Quite a couple of those issues were kernel issues. The Mac kernel is far from 
complete, so some things are expected to break. None of the things you
described were fatal. Things will improve over time, or so I hope.


TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-68k-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: