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

Re: Another installation failure report



> Problems:
> * Isn't it a bit odd to prompt for the debconf priority before prompting for 
> the language? How can I read the debconf priority prompt if I don't speak 
> english? The first prompt should be the language prompt.

Yes it's bad. I don't know a clean way to do it though, since
languagechooser should depend on cdebconf-udeb, and what happens when
you pick languagechooser from the menu is that packages are recursively
configured.

> * I'm using the newt frontend. It works pretty well most of the time. I still 
> haven't figured out how you get the cursor onto the 'go back' button in many 
> of dialog boxes.

You can tab there when the backend supports backup. :) If a package
doesn't support backup, IMO it's a bug and should be reported (including
a patch would be great :) I haven't yet figured out how to mark the
button in a different colour when it's inactive. :(

> * Hardware detection is broken. I'm using discover2 debs hacked into the 
> build, but naturally hw-detect doesn't talk to discover2 properly anyway.

What do you mean "naturally"? It doesn't work, no, since discover2
doesn't have a --format flag, and there's no "paste" program available.

> * When doing the 'loading installer modules' stage, even though I select 
> nothing in the list, a lot of module that i really don't need are loaded. I 
> really don't need,eg. the irda kernel modules for an install. Not even a 
> little bit. Is this because they are of Priority: standard or something?

Yes. A future project of mine is to change that question so that
packages with priority standard will appear in the list too, with
checked boxes. But this question will most likely be asked at priority
low, as only a nitpicker should need to answer it.

> * fdisk is not user friendly. cfdisk is in the image, so why didn't the 
> installer use it instead of fdisk in the 'Partition a hard disk'? Why instead 
> is it in 'Partition a hard disk (OLD)'? Can we kill one or both from the 
> installer image?

partitioner (the new udeb) and di-utils-partitioner (the "OLD" udeb)
both have code that checks which frontend is running (and which arch
we're on) and tries to launch an appropriate partitioning program.
partitioner launches cfdisk if the frontend is "slang", but it should
definitely do so for "newt" too. Fixing this in CVS right now.

> * Debootstrap doesn't actually appear to output its progress via the frontend 
> yet, it's just showing a bunch of I: and P: messages on stdout, which is less 
> than pretty.

These are parsed by boot-floppies, and we will parse them too,
eventually, but there's the problem that cdebconf progress bars use
templates so the texts can be translated, whereas debootstrap just spits
out English text...

> * Debootstrap failed because console-tools depends on libconsole, which 
> doesn't exist (#195766). Unfortunately, the front-end didn't print this 
> information out for me - I had to go grubbing in the log file myself. It 
> might be good if we could display the error log for the user on debootstrap 
> failure, or something like that.

I didn't have any problems with debootstrap when I did a test run a few
hours ago... Strange. Anyway, please file a bug on base-installer for
not interpreting debootstrap's error messages.

> * If I execute a shell from the installer menu, once I exit the shell for some 
> reason the vga framebuffer module is loaded, which is actually really 
> annoying (it screws up the screen). It isn't loaded by default - why does 
> closing a shell load it?

Uh. You got me there. Sounds like something crashed and init restarted
everything (rootskel tries to load framebuffer modules). Hard to know
what went wrong...


Thanks for the report.


/Martin
-- 
Martin Sjögren
  sjogren@debian.org -- marvin@dum.chalmers.se
  GPG key: http://www.mdstud.chalmers.se/~md9ms/gpg.html
  let hello = "hello" : hello in putStr (unlines hello)

Attachment: signature.asc
Description: Detta =?ISO-8859-1?Q?=E4r?= en digitalt signerad meddelandedel


Reply to: