Re: woody boot-floppies 3.0.9 i386 for testing
Eduard Bloch <edi@gmx.de> writes:
> In addition, some thinks I would like to change (some extracted changes
> from my ext3 BF set):
I'd like to put this off to after I get 3.0.9 tagged and released.
> Why this '[^a-zA-Z]'? This make the script ignore my "-Custom*" version
> of the kernel:
>
> --- boot-floppies-3.0.8.orig/common.sh
> +++ boot-floppies-3.0.8/common.sh
> @@ -91,7 +91,7 @@
> # package name might have '=' or /', so get rid of that
> pkg=${pkg%%[/=]*}
> # force files in the 'updates' dir to alway override
> - for j in `pwd`/updates/${pkg}_[^a-zA-Z]*.deb; do
> + for j in `pwd`/updates/${pkg}_*.deb; do
> if [ -f "$j" ]; then
> f="$f $j"
> fi
Ok.
> We could increase the size of the ramdisk, I did not get any problems
> with 4MB:
Just because we can does not mean we should. Raising this increases
the memory requirements.
If we get the library reduction stuff improved, there should be plenty
of space to play with.
So -- no approved.
> This is weird... As far as I can see, the function is_fstype returns the
> value of strncmp. But strncmp returns 0 for equal no 1/-1 for not equal.
> So this should be changed (the author of baseconfig.c may correct this):
>
> --- boot-floppies-3.0.8.orig/utilities/dbootstrap/baseconfig.c
> +++ boot-floppies-3.0.8/utilities/dbootstrap/baseconfig.c
> @@ -250,7 +250,7 @@
> else {
> /* this no longer seems too intelligent? */
> status = is_fstype ("/target", "reiserfs");
> - if ( status == -1 )
> + if ( status != 0)
> {
> perrorBox (_("Could not determine FS type of /, assuming ext2"));
> status = 0; /* no match, thus ext2 */
No, it returns -1 for error, 1 for match, and 0 for no match.
> This may be interessting for Herbert Xu too. Why do we use "msdos" as
> the default DOS-like filesystem in the kernel? It doesn't support long
> filenames, and when built-in, it is used first while auto-detecting
> fs-type. I suggest to drop msdos support from the kernel (or compile as
> module, since sometimes needed) and use vfat instead:
Indeed an interesting request. Herbert, any comments?
The other patches seem decent, but better not to mess with right now
so I can get 3.0.9 out.
--
.....Adam Di Carlo....adam@onshore.com.....<URL:http://www.onshored.com/>
Reply to: