On Mon, Mar 15, 2004 at 08:53:15PM +0100, Sven Luther wrote: > On Mon, Mar 15, 2004 at 08:38:29PM +0200, Anton Zinoviev wrote: > > On 15.III.2004 at 19:17 Sven Luther wrote: > > > > > > Well, the problem is as follows : > > > > > > 1) the -powerpc kernel used in most 32 bit powerpc installs has the ext3 > > > fs builtin. > > > > > > 2) the -powerpc-small kernel used in oldworld pmacs, but not yet > > > supported by d-i, has a ext3 module. > > > > This complicates things. partman-ext3 depends on ext3-modules and as > > a result ext3-modules will be installed even on newworld pmacs. I > > suppose that the best solution of this would be the newworld kernel > > package to declare "Provides: ext3-modules". > > This would be for the kernel .udeb, right, and so it would be fixed in > linux-kernel-di, right ? > > > > Partman wrongly supposes that ext3 is modular, and follows up into > > > installing and insmodding the -powerpc-small modules into the -powerpc > > > kernel. No idea if this succeeds or not, but this is dead wrong. > > > > If it succeeds then partman-ext3 will consider ext3 as valid file > > systems. Otherwise partman-ext3 will find ext3 in /proc/filesystems > > and as a result ext3 will be again a valid file system. I don't know > > why this is not so. > > > > (By the way partman-ext3 has this behaviour only since version 17 from > > 14 Mar 2004 11:04:29 +0200. The versions before 17 ignore > > /proc/filesystems. It is possible that you used somewhat older > > version of partman-ext3.) > > Yeah, naturally, i tested this yesterday, so ... > > I will test again and provide feedback to you tomorrow.
Yes, today ext3 is again proposed. Thanks. Friendly, Sven Luther -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
