[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Donnie Berkholz]

On 13:36 Sat 13 Oct     , Matti Bickel (mabi) wrote:
> 1.1                  sys-cluster/pvfs2/pvfs2-2.6.3-r1.ebuild
> 
> file : http://sources.gentoo.org/viewcvs.py/gentoo-x86/sys-cluster/pvfs2/pvfs2-2.6.3-r1.ebuild?rev=1.1&view=markup
> plain: http://sources.gentoo.org/viewcvs.py/gentoo-x86/sys-cluster/pvfs2/pvfs2-2.6.3-r1.ebuild?rev=1.1&content-type=text/plain

> 	if kernel_is 2 4; then
> 		BUILD_TARGETS="just_kmod24"
> 		ECONF_PARAMS="--with-kernel24=${KV_DIR}"
> 		MODULE_NAMES="pvfs2(fs::src/kernel/linux-2.4)"

Feel free to drop 2.4 support.

> 	if kernel_is gt 2 6 20 ; then
> 		epatch "${FILESDIR}"/${PV}-register_sysctl_table.patch
> 	fi
> 
> 	if kernel_is ge 2 6 22 ; then
> 		epatch "${FILESDIR}"/${PV}-kmem-and-dtor-fix.patch
> 	fi

Mixing 'gt' and 'ge' is a bad idea.

Thanks,
Donnie
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Drake Wyrm]

[-- Attachment #1: Type: text/plain, Size: 599 bytes --]

Donnie Berkholz <dberkholz@gentoo.org> wrote:
> On 13:36 Sat 13 Oct     , Matti Bickel (mabi) wrote:
> > 	if kernel_is gt 2 6 20 ; then
> > 		epatch "${FILESDIR}"/${PV}-register_sysctl_table.patch
> > 	fi
> > 
> > 	if kernel_is ge 2 6 22 ; then
> > 		epatch "${FILESDIR}"/${PV}-kmem-and-dtor-fix.patch
> > 	fi
> 
> Mixing 'gt' and 'ge' is a bad idea.

Just outa curiosity, why?

-- 
Each night Father fills me with dread
When he sits on the foot of my bed.
        I'd not mind that he speaks; In gibbers and squeaks,
But for the seventeen years he's been dead.  --  Edward Gorey

[-- Attachment #2: Type: application/pgp-signature, Size: 206 bytes --]

Re: [gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Donnie Berkholz]

On 23:45 Sat 13 Oct     , Drake Wyrm wrote:
> Donnie Berkholz <dberkholz@gentoo.org> wrote:
> > On 13:36 Sat 13 Oct     , Matti Bickel (mabi) wrote:
> > > 	if kernel_is gt 2 6 20 ; then
> > > 		epatch "${FILESDIR}"/${PV}-register_sysctl_table.patch
> > > 	fi
> > > 
> > > 	if kernel_is ge 2 6 22 ; then
> > > 		epatch "${FILESDIR}"/${PV}-kmem-and-dtor-fix.patch
> > > 	fi
> > 
> > Mixing 'gt' and 'ge' is a bad idea.
> 
> Just outa curiosity, why?

Because it's inconsistent and one generally assumes that people will be 
consistent with the way they test numbers. That way you only need to 
read the number rather than continually checking every single line to 
see how exactly it's tested for.

Thanks,
Donnie
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Steve Long]

Donnie Berkholz wrote:
> On 23:45 Sat 13 Oct     , Drake Wyrm wrote:
>> Donnie Berkholz <dberkholz@gentoo.org> wrote:
>> > On 13:36 Sat 13 Oct     , Matti Bickel (mabi) wrote:
>> > > if kernel_is gt 2 6 20 ; then
>> > > epatch "${FILESDIR}"/${PV}-register_sysctl_table.patch
>> > > fi
>> > > 
>> > > if kernel_is ge 2 6 22 ; then
>> > > epatch "${FILESDIR}"/${PV}-kmem-and-dtor-fix.patch
>> > > fi
>> > 
>> > Mixing 'gt' and 'ge' is a bad idea.
>> 
>> Just outa curiosity, why?
> 
> Because it's inconsistent and one generally assumes that people will be
> consistent with the way they test numbers. That way you only need to
> read the number rather than continually checking every single line to
> see how exactly it's tested for.
> 
I don't see how this is inconsistent either: two tests are needed, so that
both patches are only applied for >=2.6.22 and first only if >2.6.20.

(If the eclass performs inconsistently, that would need to be fixed.)


-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Matti Bickel]

[-- Attachment #1: Type: text/plain, Size: 833 bytes --]

Steve Long <slong@rathaus.eclipse.co.uk> wrote:
> >> > Mixing 'gt' and 'ge' is a bad idea.
> >> 
> >> Just outa curiosity, why?
> > 
> > Because it's inconsistent and one generally assumes that people will be
> > consistent with the way they test numbers. That way you only need to
> > read the number rather than continually checking every single line to
> > see how exactly it's tested for.
> > 
> I don't see how this is inconsistent either: two tests are needed, so that
> both patches are only applied for >=2.6.22 and first only if >2.6.20.

The point is that if you stick to "ge" OR "gt", everyone could just skip
reading the comparison and focus on the numbers. Will be fixed in the
next release, along with kernel-2.4 support...

-- 
Regards, Matti Bickel
Signed/Encrypted email preferred (key 4849EC6C)

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-dev] Re: Re: Re: [gentoo-commits] gentoo-x86 commit in sys-cluster/pvfs2: ChangeLog pvfs2-2.6.3-r1.ebuild

[Steve Long]

Matti Bickel wrote:

> Steve Long <slong@rathaus.eclipse.co.uk> wrote:
>> >> > Mixing 'gt' and 'ge' is a bad idea.
>> >> 
>> >> Just outa curiosity, why?
>> > 
>> > Because it's inconsistent and one generally assumes that people will be
>> > consistent with the way they test numbers. That way you only need to
>> > read the number rather than continually checking every single line to
>> > see how exactly it's tested for.
>> > 
>> I don't see how this is inconsistent either: two tests are needed, so
>> that both patches are only applied for >=2.6.22 and first only if
>> >2.6.20.
> 
> The point is that if you stick to "ge" OR "gt", everyone could just skip
> reading the comparison and focus on the numbers. Will be fixed in the
> next release, along with kernel-2.4 support...
> 
OIC: so the argument was it should be ge 2.6.21 as well? Does that catch all
the same cases?

I must say I find this criticism unusual: if someone were looking at the
ebuild to check the numbers, I would guess it were because something was
going wrong. As such, they would be paying attention to which version they
were on, and what the tests were. I don't see the use-case for limiting
what maintainers can do in such a fashion, but if it makes no difference to
the outcome (ie which cases are covered), i guess it makes sense.


-- 
gentoo-dev@gentoo.org mailing list