[PC-BSD Testing] Testing Digest, Vol 8, Issue 35
kris at pcbsd.com
Thu Aug 30 08:36:17 PDT 2007
Terry Poulin wrote:
>> Message: 2
>> Date: Tue, 28 Aug 2007 18:07:54 -0700 (PDT)
>> From: sc at sdf.lonestar.org
>> Subject: [PC-BSD Testing] Add nano/joe editors, and a little bootup
>> To: testing at lists.pcbsd.org
>> Message-ID: <642188.8.131.52.44.1188349674.squirrel at sc.freeshell.org>
>> Content-Type: text/plain;charset=iso-8859-1
>> Kris et al.,
>> This pertains to last Friday's snapshot (and earlier), but since I haven't
>> heard anyone else mention this, I think it would be nice to offer a choice
>> of a few widely familiar and comfortable text editors right from the
>> start, that could be available when things go badly during or after an
>> The included "edit" editor is OK, but many people probably could work
>> better and make fewer mistakes with joe or even pico/nano.
> Any thing simple and with on screen guides should work, but Pico would
> probably be best... It's as basic as ed yet as simple as notepad in a console
> ready package.
I can concede to adding "pico" to the base, I have a soft spot for it,
since its the editor I first learned FreeBSD on, before having to master
"vi" :) That'll probably help us with troubleshooting when a user needs
to edit a file from the command-line.
>> Also, the bootup process is still screechingly at odds with PC-BSD's great
>> work elsewhere at making FreeBSD easier to use. When an install didn't
>> work recently because an old non-booting disk drive that was attached to
>> the system "failed" (as far as I could see, it didn't pass fsck because of
>> 1 bad cluster), all the wonderful GUIs and goodness of PC-BSD vanished
>> when the bootup process tossed itself overboard at a point when not even
>> joe could run because the state of things yet was such that random cursor
>> addressing for the screen wasn't working.
>> The system behaved as if this disk was critical to its operation. That was
>> far from true, and an easy-to-use option to bypass it rather than
>> struggling against the system's wrong assumptions would have been very
>> welcome. Some errors can be addressed much better after the boot disk and
>> the system are fully up and operational. Forcing the user to solve such
>> problems at a partially effective system prompt with severely limited
>> tools is sub-optimal. More than once while this was happening, not only
>> that ancient disk but the main boot disk too were marked read-only and not
>> even root could edit the necessary files to get around it.
>> Some useful messages (understandable to the average person) about how to
>> bypass this kind of thing are fairly nonexistent in FreeBSD's bootup
>> process. It would be very helpful to look through the bootup procedure to
>> see what helpful informational messages and screens screens could be added
>> at various points, to help new desktop users who aren't Unix gurus when
>> bootup falls over and throws PC-BSD on its ear in the mud.
>> At least, giving the user clearer ways to differentiate between really
>> showstopping errors and lesser errors that could be fixed more effectively
>> once the desktop GUI system is up, compared to trying to fix things
>> partway through bootup using sticks and flint knives, seems desirable if
>> at all possible.
> I suppose that is a question of how true to the underlaying FreeBSD the
> developers wish to stay with what changes they do make.
As for the boot process, these are all valid points. I'm going to put
this down to investigate for the next release. We'll just have to figure
out if we can add some sort of "monitoring" service, which keeps track
of the startup, and takes appropriate actions if the load fails.
More information about the Testing