bug-parted
[Top][All Lists]
Advanced

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

Re: parted pre9


From: Andrew Clausen
Subject: Re: parted pre9
Date: Sat, 11 Nov 2000 04:42:50 +1100

Hi Ethan,

> when i tried to create an extended partition on a pc98 label i got:
> 
> (parted) mkpart extended 100 150
> You found a bug in GNU Parted.  Please email a bug report to
> address@hidden containing the version (1.4.0-pre9), and the
> following message:
> 
> Assertion (supports_extended || type != PED_PARTITION_EXTENDED) at
> disk.c:615 in function ped_partition_new() failed.
> Ignore Cancel ? c
> (parted)
> 
> i assume pc98 does away with the primary/logical/extended sillyness?

Yes.  I just added a prettier error message, BTW.

> also it seems that every partition you create in a pc98 label is
> marked bootable, and cannot be marked otherwise is this normal?

Yes.

> i also noticed when you create a intel partition table parted asks if
> you want to install the parted bootloader (if none exists) if you say
> no, parted will ask again on your next command (mkpart whatever) and
> if you still say no it will ask you again on the next command (set 2
> boot on whatever) iow it will ask you over and over and over and over
> and over and over again until you give in and say yes ;-)

Yes, I know!  I can't think of a good workaround.

When we merge with partimage, there should be no need for the
bootstrap code (I suspect...), so this will go away, eventually...

> i can't reproduce the spurious Apple_Void creation anymore.

They still get created!  BTW, corezion tested it with MacOS.  MacOS
likes it :-)

> can you do anything to make `mkpart primary hfs 0.032 0.812' create a
> 1600 block partition instead of 1599?  (0.813 creates it a bit over
> 1600)

Yes.  s/0.812/0.8125/

> one limitation i found in the regular intel partition table editing,
> you cannot set the partition type with parted, it seems to try and set
> it for you based in the filesystem you choose but this won't always
> work.

When won't it work?

BTW: it is based on the file system AND the partition flags.

BTW2: it's MUCH better this way.  99.9% of users don't know which
partition type they should select (even those who THINK they do!)
The flags system makes it much more intuitive (and looks cute
in GUIs ;-)

> one albeit contrived example is `mkpart primary|logical hfs 20 30'
> which sets the partition type to 83.

Yes.  What's wrong with this?

> however a less contrived example
> is RS6000 powerpcs or PReP boxes, which use intel partition tables and
> use a type 41 PReP Boot bootstrap partition, where a bootloader or the
> kernel is dded directly on it.  (or some others use a FAT filesystem
> with a bootloader similar to the NewWorld ybin/yaboot setup)  there is
> also BSD disklabel partitions, type a6 for OpenBSD a7 for NeXTSTEP,
> and many many others.   i think parted really need to display the
> partition type (not just primary,logical,extended) and allow for it to
> be explicity changed.

This support can (and will, if ppl scream/care enough to send in
patches) be added.  However, I can't see any reason to move away
from the FS & flag system.

> otherwise people will have to use the old fdisk
> just to fix partition types.

Not when we support all the (useful) types.  I don't think we need
to support CP/M  *grin*

> a way to set arbitrary types on mac
> partitions would be useful too IMO.

Why?

> can you change `set x boot on' to set the partition name to
> `bootstrap' for mac partition tables?  i think this is better then
> leaving it `untitled'.

What if the user (or MacOS) gave it a different name?
 
> one idea i had about the interface of parted, what about making some
> commands optionally interactive?  so instead of having to type `mkpart
> primary blah foo bar' over and over again, run:
> 
> mkpart
> 
> which then prompts for primary or logical, with default being primary
> (or logical if thats all that is possible anymore).  then it asks for
> filesystem type, default ext2, then it asks start, then end.
> 
> something like:
> 
> (parted) mkpart
> (parted) mkpart> type? [primary]:
> (parted) mkpart> filesystem? [ext2]:
> (parted) mkpart> start? [5.001]:
> (parted) mkpart> end? [258.213]:
> (parted)
> 
> just a rough idea...

Yep.  This idea has already been suggested.  Maybe, One Of These
Days, but there are so many more interesting/important things to
do!

> one could also run `mkpart --batch' where once a partition is created
> it immediatly starts over for the next partition, until no more space
> is left.  mkpartfs could work identicly except for formatting the
> partition.

Don't like this idea.  Too complicated.  I prefer your "-" idea.

> this would save for alot of typing and allow for a similar (but far
> better) interface to fdisk.  i personally find the interactive
> partition creation of fdisk to be quite convenient and allows me to
> partition a disk very quickly.

hehe, except it's broken when you make a mistake.  You have to start
all over again.

> it would also be more consistent with
> the way parted itself works, as in you can run either:
> 
> parted /dev/sda mkpart primary ext2 1 5
> 
> or
> 
> parted /dev/sda
> > mkpart primary ext2 1 5
> > quit
> 
> what do you think?  (not for 1.4 but maybe in the next version)

Ah.  I see.  Hmmm.  If we had a proper lexer / parser, it would
be quite cute... One Of These Days TM.

Have to think about the interface a bit though.  (Feedback WRT
which partition you're up to, etc.)  All you Mac people have to
Think Different TM with your UIs, don't you!  :p
 
> also you have my email address in the THANKS file but its wrong, im
> address@hidden not address@hidden (unfortuantly, someone else got
> `eb' ...) thanks for the mention though ;-)

Fixed :-)

Thanks for your comments :-)

Andrew Clausen



reply via email to

[Prev in Thread] Current Thread [Next in Thread]